Division of Machine Design TQFT33 | Master Thesis
Master Thesis Aeronautical Engineering
Spring term 2019
LIU-IEI-TEK-A–19/03450—SE
On-board Data Acquisition System
Conceptual Design of an Airdrop Tracking System
Hanna Eriksson
Academic supervisor: Jonas Detterfelt
External supervisor: Martin Svensson (Saab AB) Examiner: Johan Ölvander
Linköping universitet SE-581 83 Linköping, Sverige
This thesis is, on behalf of Saab AB, a pre-study of possible on-board solutions for position measuring during store separation tests aimed for the test and evaluation of JAS 39 Gripen. The purpose is to replace the present ground-based system in order to achieve more effective trials regarding time and economy.
Three different concept development methodologies were investigated in order to find the most suitable one for this thesis. Those were merged into one adapted methodology containing the following phases; Planning, Function Analysis, Concept Generation and Concept Evaluation.
The work progressed as the methodology states, and the highest amount of work was dedicated to the Planning phase. The requirements and desiderata for the system were produced with an agile process, resulting in the Construction Specification List that eventually became the basis for the Concept Generation phase.
Knowledge about the technical theory needed to solve the problem was obtained in parallel with the Function Analysis and Concept Generation. The most adaptable techniques to measure position were found out to be with the use of the Global Positioning System (GPS) or Inertial Navigation System (INS).
After an extensive work with the Concept Generation in parallel with a continuously updated Construction Specification List, three concepts were developed. One concept is based on GPS, the second one on INS and the third one is a combination of GPS and INS. All three concepts shares the same telemetry system and casing, which fulfills the requirement of simple installation and possibility to install in different stores.
In the final phase, Concept Evaluation, a comparison between the concepts was performed. Advantages and disadvantages was listed and the fulfillment of requirements was investigated. All three concepts were handed over to Saab in order to let them decide which concept(s) to further develop.
1 A Cinetheodolite used at Denel Overberg Test Range [1] . . . 5
2 The INS 65210E manufactured by Measurement Specialities [2] . . . 6
3 The MTi-G-710 system manufactured by Xsens Technologies [3] . . . 7
4 The concept development process according to Ulf Liedholm . . . 9
5 The Function Analysis according to Liedholm . . . 10
6 The five-step concept generation method according to Ulrich and Eppinger [4] 11 7 Concept generation methods according to Ullman . . . 13
8 Morphological matrix according to Ullman [5] . . . 14
9 Example of an evaluation matrix according to Liedholm [6] . . . 15
10 Example of a concept screening matrix according to Ulrich and Eppinger [4] 16 11 The concept scoring matrix according to Ulrich and Eppinger [4] . . . 17
12 Example of a decision matrix according to Ullman [5] . . . 19
13 The 24-satellite constellation in six orbital planes [7] . . . 22
14 A schematic view of the control segment elements and there correlation [7] . 23 15 Fundamentals of Trilateration[8] . . . 23
16 Intersection between different satellites [8] . . . 23
17 Principle of real-time Kinematic [9] . . . 25
18 The definition of body and global frames in relation to the object [10] . . . 26
19 Example of a stable platform IMU [10] . . . 26
20 The inertial navigation algorithm for a Stable Platform System [10] . . . 26
21 Example of a Strap-down IMU [11] . . . 27
22 The inertial navigation algorithm for a Strap-down System [10] . . . 27
23 A typical light weight accelerometer [12] . . . 28
24 Example of a mechanical gyroscope [13] . . . 30
25 Visualization of the Sagnac Effect. The dashed line shows the path taken by the beam travelling with the direction of rotation [10] . . . 30
26 Coriolis effect when angular velocity is applied [14] . . . 31
27 Summation of the time and measurement update equations used in the Extended Kalman Filter[15] . . . 33
28 Schematic picture of a telemetry system [16] . . . 34
29 Comparison of antenna types . . . 36
30 The frequency bands defined by ITU. The range applicable in this thesis is circled with green . . . 36
31 Methodology adapted for this thesis . . . 37
32 Content of the Planning phase . . . 38
33 Content of the Function Analysis . . . 39
34 Content of the Concept Generation phase . . . 40
35 Content of the Concept Evaluation . . . 41
36 Construction Specification List . . . 43
37 Black-Box model of the system . . . 44
38 Function/Mean-tree for the system . . . 44
39 Functional flow of Concept 1 . . . 45
40 Functional flow of Concept 2 . . . 47
1 Introduction 1 1.1 Background . . . 1 1.2 Objective . . . 1 1.2.1 Deliverables . . . 1 1.3 Problem description . . . 2 1.3.1 Question formulations . . . 2 1.3.2 Delimitations . . . 2 1.3.3 Requirements . . . 2 1.3.4 Desiderata . . . 3 1.4 Planning . . . 3 1.5 Method . . . 3 2 Situation assessment 5 2.1 Present solution . . . 5 2.2 Market analysis . . . 5 2.2.1 TE Connectivity-Measurement Specialities . . . 6 2.2.2 Xsens Technologies . . . 7 3 Theory of Methodology 9 3.1 Concept Generation . . . 9
3.1.1 Concept generation according to Liedholm . . . 9
3.1.2 Concept generation according to Ulrich & Eppinger . . . 11
3.1.3 Concept generation according to Ullman . . . 12
3.2 Concept Evaluation and Selection . . . 15
3.2.1 Concept evaluation and selection according to Liedholm . . . 15
3.2.2 Concept evaluation and selection according to Ulrich & Eppinger . . 16
3.2.3 Concept evaluation and selection according to Ullman . . . 18
4 Theoretical frame of reference 21 4.1 Ballistics . . . 21
4.2 Position Measuring . . . 21
4.2.1 Global Positioning System (GPS) . . . 21
4.2.2 Differential GPS . . . 24
4.2.3 Real-Time Kinematic (RTK) . . . 24
4.2.4 Inertial Navigation System (INS) . . . 25
4.2.5 Combined GPS/INS with Kalman Filter . . . 31
4.3 Data Acquisition . . . 34 4.3.1 Computer storage . . . 34 4.3.2 Data Transmission . . . 34 5 Methodology 37 5.1 Planning . . . 38 5.1.1 Define Task . . . 38 5.1.2 Situation Assessment . . . 38
5.3 Concept Generation . . . 39
5.3.1 Combine means . . . 40
5.3.2 Externally & internally search . . . 40
5.4 Concept Evaluation . . . 41
5.4.1 Check if requirements are fulfilled . . . 41
5.4.2 List advantages and disadvantages . . . 41
6 Results 43 6.1 Planning . . . 43 6.2 Function Analysis . . . 44 6.3 Concept Generation . . . 45 6.3.1 Concept 1 - GPS . . . 45 6.3.2 Concept 2 - INS . . . 47 6.3.3 Concept 3 - GPS/INS . . . 48
6.3.4 Conceptual Design of the System Casing . . . 49
6.4 Concept Evaluation . . . 50 7 Discussion 53 7.1 Discussion of Methodology . . . 53 7.1.1 Planning . . . 53 7.1.2 Function Analysis . . . 53 7.1.3 Concept Generation . . . 54 7.1.4 Concept Evaluation . . . 54 7.2 Discussion of results . . . 54 8 Conclusion 57 8.1 Recommendation of future work . . . 57
Appendices 63
1
Introduction
This thesis is a pre-study of possible on-board solutions for position measuring during store separation tests at Saab AB.
1.1
Background
Saab AB is a Swedish aerospace and defence company founded in 1937 [17]. One of the six business areas, Aeronautics, engages in advanced development of military and civil aviation technology. The most known and extensive product is the JAS 39 Gripen fighter. Gripen is a single-engine multirole fighter aircraft with a delta wing and canard configuration.
The hardware and software in the aircraft is continuously updated and there is always a new version in progress. During the development of a new aircraft, test and verification is an important part. The Flight Test and Verification department is responsible for test and evaluation of prototypes, modified aircraft and production aircraft through ground-, simulator and flight testing.
One part of the development process of a new fighter, is the integration and test of external stores. Separation tests are an important part of the integration phase, both as a verification of a safe separation and as a way to investigate the ballistic trajectory of the store until it reaches the target. At the moment, these tests are performed at a specific test area that provides advanced, stationary measuring systems. The fact that the test area has to be reserved far in advance makes it a narrow resource. Also, the investigation and evaluation of the store’s ballistic trajectory is very time-consuming with the present system.
1.2
Objective
The objective of this thesis is to investigate the possibilities of replacing the present ground based measuring system, used in external store separation trials, with an on-board system. The purpose with this substitution is to enable a more flexible separation trial process that is more time efficient and can be performed independent on the resources available at the test range.
1.2.1
Deliverables
The project should produce the following deliverables: • A report
• Possible and impossible solutions • Recommendation of future work
1.3
Problem description
Separation tests of external stores are today performed at a specific test area with advanced measuring systems provided. These systems are stationary and requires manual operation. The test area is a narrow resource and today’s test campaigns requires a large amount of resources. Due to these factors, the Tactical Systems and Integration section at Saab has a desire to investigate the possibilities of an on-board system attached directly on or inside the external store, and evaluate in which way this system could replace the present one.
1.3.1
Question formulations
Since the thesis is a concept development project, there are no specific question formulations to answer. Though, two general questions are formulated to aid the line of argument in the thesis.
• Given the limitations, requirements and desiderata, is there one or several concepts that fulfill those?
• What limitations does each concept have?
1.3.2
Delimitations
In order to assure a valuable outcome of the project within the given time frame, some delimitations was set:
• The external store is an XXXX, unguided bomb
• The cost for material and manufacturing must not exceed 40 000 SEK per system unit.
1.3.3
Requirements
The specification of requirements is minimal to enable a higher amount of possible solutions. These initial requirements set by Saab, are going to be more specified and extended later in the design process:
• The installation must not have any apparent influence on the separation or the trajectory of the load
• The product must withstand forces up to 100G without failure • The data collected must be saved
• The system must be reliable regarding function, accuracy and construction • It must be able to test stand alone to verify the functionality of each unit
1.3.4
Desiderata
Some functions are desired to fulfill in order to create value for the business. Though, these must not be fulfilled in order to produce a usable product.
• It should be able to monitor the data in real-time, using telemetry
• If the above mentioned function is not fulfilled, it should be possible for the pilot to see if the system works or not, before the separation is performed
• Simple installation and possibilities to install the system in other stores than the one mentioned in section 1.3.1.
1.4
Planning
The total available time for this master thesis is 20 weeks, which corresponds to 800 hours. To make a project plan, the project was divided into different phases. An initial estimate of how much time each phase requires was made, presented below:
• Pre-study and get familiar with the problem scope: 30 h • Planning, including planning report: 40 h
• Literature study, find a concept development method: 20 h • Theoretical study: 50 h
• Concept development and evaluation: 250 h • Master thesis report: 400 h
A Gantt-schedule, see Appendix A, was made in order to structure the work week by week.
1.5
Method
This project should be performed by using a concept development method. As a beginning, existing solutions should be investigated along with an update of the requirement specification list. Then a research of existing concept development methods should be performed, and the most usable method should be chosen. Using the chosen method, concepts should be generated and the process iterated until possible and impossible solutions are found. The concepts should then be evaluated.
2
Situation assessment
The techniques required by the system to accomplish the requested measurements is already used in many similar applications. For example, aircraft always has one or several different systems on-board that continuously measures its position and altitude. Those systems are usually based on the Global Positioning System or Inertial Navigation Systems. In this section, an analysis of the current situation is presented. The present ground-based solution that Saab uses for these trials today are also investigated here.
2.1
Present solution
Currently the ballistic trajectory of a store during a separation test is measured with a ground system based on Cinetheodolites. A Cinetheodolite is a photographic instrument for measurement and collection of trajectory data. The system can provide measurements of position and event data. [18] It uses the fundamentals of motion pictures, where the cameras are mounted with its optical axis free to move in all directions. This enables the camera to always be aimed against the object, usually a missile or aircraft. [1]
The Cinetheodolite records the elevation angle, azimuth angle and the time of each frame of film exposed during the tracking of an object. The film from two or more theodolites can then be used to determine the ballistic trajectory of the object. For three-dimensional tracking, at least three theodolites are required, while two is enough for two-dimensional tracking.
Figure 1 shows an example of a Cinetheodolite used for aerospace related testing at Denel Overberg Test Range in South Africa. [1]
Figure 1: A Cinetheodolite used at Denel Overberg Test Range [1]
2.2
Market analysis
There are already several existing solutions that can meet the requirements for the on-board data acquisition system requested in this thesis. However, they cost too much to be valuable for using in separation trials, since each unit of the system only can be used once.
2.2.1
TE Connectivity-Measurement Specialities
An American company named TE Connectivity-Measurement Specialities, manufactures inertial measurement systems for several different applications. They have one system that is adapted especially for use in a unguided Mk-80 series bomb; [2]
User Configurable Inertial Measurement System-65210E
This is a six degrees of freedom INS which contains the following components: • three internal accelerometers
• three internal rate gyros • two temperature sensors
• battery voltage and current monitor • signal processor
• IRIG encoder
• optional FM transmitter • high-capacity Li-Ion battery
Figure 2: The INS 65210E manufactured by Measurement Specialities [2]
All of these components are installed in a small cylindrical package that will fit a standard Mk-80 fuse, see Figure 2. All channels in the system are continuously measured, each sampled at 16 bits. The data is then filtered, ranged and calibrated at 42 500 samples per second and channel. The system is suitable for harsh environments since it can operate in a temperature range between -40 to +85 degrees Celsius, and will sustain a shock of 100G. The system also contains a telemetry transmitter that can meet the desire of real-time data monitoring. This system fulfills each requirement specified for the desired On-board Data Acquisition System that this thesis aims to develop, but the cost is too high. [2]
2.2.2
Xsens Technologies
Xsens is a Dutch company that develops 3D motion tracking technologies and products. They develop Attitude and Heading Reference Systems (AHRS), which provides 3D orientation data by fusing the integrated gyroscope data with data from accelerometers and magnetometers. The aim of the fusion is to reduce the drift from the gyroscopes by compensating with the magnetic field of Earth. [3]
MTi-G-710
The MTi-G-710 is a miniature AHRS based on both GPS and INS, that provides high-quality position, velocity, acceleration and orientation. The system has excellent heading tracking without requiring a magnetic field. The system can operate in rough and challenging environments and can sustain a shock of 2000G for 0,5 ms. The system consists of gyros, accelerometers and magnetometers. This system does not include any transmitter for telemetry of data in real-time. [3]
3
Theory of Methodology
Product development is a complex process, which can be done using several different methods. The process reaches from the first idea until the final physical product. One important stage of the process is the concept development. In this chapter, some different concept development methodologies are explained shortly and later compared with each other.
3.1
Concept Generation
The first step of the concept development phase is the generation of possible solutions. The concept generation can in turn consist of several phases depending on the method used. In this section, three different concept generation methodologies are presented.
3.1.1
Concept generation according to Liedholm
The concept generation process consists, according to Liedholm, of three phases; Construction Specification List, Function Analysis and Concept Establishment [6], showed in Figure 4.
Figure 4: The concept development process according to Ulf Liedholm
In the first step, an investigation of the problem and a break-down of specifications should be performed. It is, according to Liedholm, important that the construction specifications are solution independent, in order to hold most possibilities open early in the process. The next step is to specify which functions the product should have and investigate possible means and solutions. The output from this stage is a function/mean-tree, which is a structured way to show the functions and the alternative means to implement them. Figure 5 shows a flow chart describing the Function Analysis process. [6]
In order to define the main function, a Black-Box model should be drawn up. The technical principles are then divided into transformation systems, which defines the functions needed to enable the main function. In the Function/Mean-tree, each technical principle is then provided with sub-functions. For each function or sub-function, one or several means are defined. A mean is a method of how to realize the function. [6]
Figure 5: The Function Analysis according to Liedholm
The third and last step of this concept development process is the establishment of concepts. The concepts are created by combining the appropriate means with each other, to solve the functions stated in the Function/Mean-tree. This process should be iterated until enough useful concept have been found.
3.1.2
Concept generation according to Ulrich & Eppinger
The concept generation process consists, according to Ulrich and Eppinger, of five different phases. The idea behind this approach is to break a complex problem into simpler sub-problems. The five-step concept generation method is presented in Figure 6. [4]Figure 6: The five-step concept generation method according to Ulrich and Eppinger [4]
At first, a general understanding of the problem should be developed, and if necessary the problem should be divided into sub-problems. An ideal input to the concept generation phase is a preliminary specification list where also the customer needs have been identified and taken into consideration. Possible concepts are then identified by internal and external search procedures. The external search aims to find existing solutions to the overall problem and the sub-problems. This research involves for example interviews with users, consulting experts and patent searching. The output from this step is both existing concepts and new concepts. [4]
In step four, classification trees and combination tables are used to explore the concepts and to integrate the solution for each sub-problem to a total solution. A classification tree helps to refine the problem decomposition even more, and could for example lead to adding of a new sub-function. [4]
The last step is to reflect over the solutions and the overall process. Identify how valid and useful the concepts are and what could be improved. Despite that this process is presented in a linear manner, Ulrich and Eppinger states the importance of an iterative concept development process. [4]
3.1.3
Concept generation according to Ullman
According to Ullman, the first part of the concept generation process is to get a deep understanding of the function. To understand the function of existing devices are important and a good practise is to do a benchmark. A good source of ideas are patent literature. However, it can be rough to find the relevant information when searching for patents. [19]
The main function should be divided into sub-functions in order to get a deep understanding of the product. Ullman presents a technique for functional modelling, which is very useful when developing new products. It consists of 4 steps;
Step 1: Find the overall function that needs to be accomplished
This step aims to state the overall function, based on the customer requirements. All design problems has according to Ullman, one or two functions that are most important, and those should be reduced to one simple statement and implemented in a Black-Box. The Black-Box is based on the conservation of mass and energy principle, which means that all energy and mass that goes into the system must either come out or be stored within system. In this step, also all interfacing objects and fixed parameters must be identified. The designer should also identify important information flows, by for example answering the question "How will the customer know if the system is performing?". [19]
Step 2: Create sub-function descriptions
This step aims to break down the overall function and identify all the sub-functions needed. Here, it is important to consider what, and not how. In this step it is important to ensure that no new components are mentioned. If so, step one must be reiterated to ensure that the specifications are complete. Let each function in the break-down represent a transformation in the flow of energy, material or information. Since a product have different operating sequences, it is a good idea to think of each function in terms of preparation, use and conclusion.
Step 3: Order the sub-functions
The aim of this Step is to order the functions generated in Step 2, to accomplish the main function defined in Step 1. It is important that the functions comes in a logical order. For example they should be arranged so that the output of one function is the input to the next one.
Step 4: Refine sub-functions
The sub-functions should now be decomposed even more, in order to see if the functions can be fulfilled by existing objects or not.
Ullman presents several different methods for concept generation, stated in Figure 7. Two of them are going to be further explained in the following sections.
Concept generation methods Basic methods Brainstorming Brainwriting Analogy
Extremes and inverses
Experts, reference books and trade journals
The morphological method
Logical methods TRIZ
Axiomatic design
Figure 7: Concept generation methods according to Ullman
3.1.3.1 Morphological Method
According to Ullman [19], the morphological method is a powerful method for idea generation. It is a useful tool to create concepts of a product that is based on sub-functions former identified. An example of a morphological matrix is shown in Figure 8. The technique consists of two steps;
Step 1: Developing concepts for each function
The objective of this step is to find as many concepts as possible that can solve each sub-function identified in the decomposition. Firstly, one should develop as many alternative functions as possible, to each already specified sub-function. Then, as many means as possible should be identified for each sub-function. If it turns out that one of the functions only corresponds to one conceptual idea, that function must be further inspected, since there is not likely that only one concept fulfills the function. For example, check that the function statement does not have any nouns telling how the function should be accomplished, since a statement like that will limit the amount of possible concepts. [19]
Step 2: Combining concepts
This step aims to combine the concepts generated in step one into overall concepts that meets the functional requirements. The concepts to each sub-function, listed in step 1, should now be combined into complete conceptual designs. One concept per sub-function are chosen, and then combined into a concept design. The problem with this method is that it probably will generate too many concepts. It also assumes that each concept only accomplish one function, which is rarely the case. Despite that, Ullman states the importance of breaking down the functions in order to get a deep understanding of the concept development. [19]
3.1.3.2 Axiomatic Design
Axiomatic Design is a system design methodology developed by Professor Nam Suh in the 1970s [19]. The aim of the method is to create a logical design process where the customer needs are transformed into functional requirements and design parameters. The axiomatic approach is based on the relationships between four design domains: customer, function, physical and process. The relation between functional requirements and design parameters
Figure 8: Morphological matrix according to Ullman [5]
can be described by the relation in Equation (1): [19]
FR = A × DP (1)
where A is the design matrix. [20] The Axiomatic Design method consists of two central axioms. The first is called the Independence Axiom and stands for the importance of maintaining the independence of functional requirements. The design is better if the functional requirements are independent of each other, since a change in a specific design parameter then will impact only one single function. Therefore, the ideal situation is to have an uncoupled design, where each functional requirement is dependent on only one design parameter. [19] For an uncoupled design, the relation turns out to be as stated in Equation (2) [20]. F R1 F R2 =X 0 0 X DP1 DP2 (2) The second best case is to have a decoupled design, which means that the design matrix is lower triangular, as in Equation (3).
F R1 F R2 =X 0 X X DP1 DP2 (3) Worst case scenario, is when each functional requirement are dependent on several design parameters. This is called coupled design, and is showed in Equation (4). [20]
F R1 F R2 =X X X X DP1 DP2 (4) In a coupled design, a change in one specific design parameter will effect all the functions, which is not the desired case. The design is much more rigid if it is uncoupled. The second axiom is the Information Axiom, which implies the importance of keeping the information content of the design as low as possible. Usually, the simplest design has the highest probability to succeed.
3.2
Concept Evaluation and Selection
The Concept Selection phase is about evaluating the concepts created during the concept generation, and select one or more concepts for further investigation and testing. Even though the amount of concepts are narrowed down, it could take several iterations of this process until the final concept(s) are chosen. The amount of concepts could even increase at a beginning, for example by creating combinations of the other concepts.
3.2.1
Concept evaluation and selection according to Liedholm
The concept selection should according to Liedholm be performed in three steps [6]: 1. Investigate whether the concepts fulfills the requirements specified in the Construction Specification List
If a concept does not fulfill the requirements, the concept is either to bad, or the requirements are too strict or badly formulated. If this happens, either the concept should be rejected or the requirements must be adjusted. [6]
2. Compare the concepts with each other by establish an evaluation matrix Liedholm states that since the knowledge about the concepts are limited at this stage, a simple evaluation method is beneficial. By using a evaluation matrix, see Figure 9, an objective comparison between the concepts is received. Though, the result should not be taken too literal since the matrix does not tell if the concept will work or not. It is only a comparison relative each other and should be seen as a guidance. [6]
3. Choose a few concepts for further development
The decision of which concepts to go further with should be based on both the evaluation matrix and the result from the previous concept review. [6]
3.2.2
Concept evaluation and selection according to Ulrich &
Eppinger
Ulrich and Eppinger presents a two-stage concept selection methodology; Concept Screening and Concept Scoring [4]. Both stages are based on a decision matrix which is used to help rate, rank and finally select the most qualified concepts. A six-step process should lead the user trough each stage;
1. Prepare the selection matrix 2. Rate the concepts
3. Rank the concepts
4. Combine and improve the concepts 5. Select one or more concepts
6. Reflect on the results
The Concept Screening phase is based on the method developed by Stuart Pugh, where the purposes are to decrease the number of concepts and to improve them. In the case of more than 12 concepts under consideration, the multi-vote technique may be used to quickly reduce the number of solutions. Multi-voting means that every member of the concept development team simultaneously votes for three to five concepts. Simply the concepts with most votes are chosen to continue with the concept screening. [4]
The input to the selection matrix, see Figure 10, are concepts and criteria, where it is important that all concepts are presented with the same level of detail. These criteria are based on both the customer needs and enterprise needs such as manufacturing cost. The concepts should be presented at the same level of detail in order to get a meaningful and condign comparison. In this stage, the criteria are still very abstract and benefits from being chosen in order to distinguish the concepts. Since all criterion stated in the selection matrix are given the same weight, it is of high importance not to list the least important criteria. Otherwise, the differences among the concepts relative the more important criteria will not be distinctly reflected in the result. [4]
All concepts are rated against a carefully chosen reference concept. The reference could be either one of the concepts under consideration, or a product available on the market.
Figure 10: Example of a concept screening matrix according to Ulrich and Eppinger [4]
as 0" or "worse than -", in relation to the chosen reference. A summation of the rates is then made, and the concepts are ranked, see Figure 10. [4]
After the screening, the result is analyzed in order to find new combinations or improvement of the present ones. Since this is an iterative process, the new combinations or improved concepts are then stated in the same selection matrix and being scored and ranked along with the other ones. [4]
In order to obtain a more condign outcome from the Concept Selection, the Concept Scoring phase is performed. Here, the rating of concepts are based on the relative importance of each criterion. The criteria former stated in the screening matrix is now provided with a weighting, in relation to a reference point. As can be seen in Figure 11, new concept combinations are available and a new reference is set.[4]
Figure 11: The concept scoring matrix according to Ulrich and Eppinger [4]
In the same way as in the screening phase, the concepts should then be rated. Ulrich and Eppinger recommends that all concepts are rated with focus on one criterion at the time. Though, in this stage it is not suitable to use a single reference concept, because of the risk for reduction of the rating scale for some of the criteria. For example, if the reference concept is the easiest one to manufacture, all of the other concepts will receive the rate 1,2 or 3 ("much worse than", "worse than" or "same as"). To overcome this problem, different reference points should be used for the various criteria. The scoring process is then finished in the same way as the screening process; combine and improve the concepts, select one or more concepts, and finally reflection of results and the process.
3.2.3
Concept evaluation and selection according to Ullman
According to Ullman, the goal of the concept selection process is to "expend the least amount of resources on deciding which concepts have the highest potential for becoming a quality product". He states that the central problem in this is to choose which concept to further develop, without knowing that much about it. At this stage, it will be less risky to refine a number of concepts before limit oneself to one of them. On the other hand, this approach will require more resources and will also lead to less investigation of each concept. [19]Ullman calls this phase of the concept development for Concept Selection, and states that it is a combination of comparison and decision making. The comparison can be made in two different ways. Either it is absolute, which means that the concepts are compared to a target set by a criterion, or it is relative, by comparing the concepts with each other. [19]
As a first step, Ullman propose that the evaluation can be made based on feasibility judgement. One of the three following reactions will emerge when looking at each concept: (1) it is not feasible and will not work; (2) it might work if something else happens; (3) it is worth considering. This method will result in a better outcome the more experience the engineer has. Before rejecting any of the ideas that seems "not feasible", the concept has to be carefully considered in order to find why it is not feasible. The concepts that are judged to be worth considering are usually the hardest ones to evaluate, which makes the engineering experience essential. [19]
When the intuition phase is done, the remaining concepts will be further evaluated in a go/no-go screening. This evaluation is based on the customer defined criteria and the technology readiness level. By reformulating the customer needs to questions, they could be answered with go or no-go. The readiness level of a product tells how mature the design is, and is a good technique to refine the concept evaluation. A technology that is not mature enough to be implemented in the design will result in an expensive and time consuming process, and the risk of a bad product as outcome is high. Ullman suggests six measures that can be applied to determine the maturity of a design: [19]
1. Are the critical parameters that control the function identified?
2. Are the safe operating latitude and sensitivity of the parameters known? 3. Have the failure modes been identified?
4. Can the technology be manufactured with known processes?
5. Does hardware exist that demonstrates positive answers to the preceding four questions?
6. Is the technology controllable throughout the product’s life cycle?
Ullman states that if the answers of these questions are negative, it is a good idea to add a consultant to the team, in order to expand the knowledge field. [19]
Further, to evaluate the concepts even more, Ullman suggests a decision-based method [19], see Figure 12. The method is simple and effective for comparison of concepts. It is based on Pugh’s method, earlier mentioned in Section 3.2.2. Each concept is scored with a weight relative to another, regarding the ability to meet the criteria set by the customer. One concept is chosen as a reference, to which the other are compared with. The method is an iterative process and new alternatives are often found. According to
Ullman, the outcome of the process is most valuable if every member of the team performs it independently, and then compares the results with each other.
4
Theoretical frame of reference
In this thesis, the bomb that should be positioned are both unmanned and unguided. Since it is not dirigible, it is of high importance to be aware of the behaviour that the bomb shows through the air and at the impact. Even though the manufacturer of the bomb guarantees the behaviour of it, Saab needs to ensure that it works as desired even after the integration to a new aircraft. At this point, the desire of an on-board system for determination of the ballistic trajectory of the bomb arises. In this section, the technical theory required to solve the problem is going to be examined and presented.
4.1
Ballistics
Ballistics is the theory about the behaviour and effects of thrown bodies or launched projectiles through the air. When analyzing thrown bodies, the effects of air resistance is often neglected and the Earth’s gravitation is considered as the only force acting on the body. This assumption gives a symmetrical parabolic trajectory through the air. However, in the real case the body is also affected by aerodynamic drag caused by the resistance of the air, resulting in an asymmetrical parabolic path with the highest point closer to its point of impact. This is especially important to consider when analyzing the motion and trajectory of a projectile. [21] The drag contribution emerged by the air varies with the velocity and altitude of the projectile. By measuring the projectile’s position along the trajectory, the ballistics of it can be calculated. The knowledge of the ballistic behaviour can then be used to adjust the rear sight of a projectile to ensure the correct point of impact. [21]
4.2
Position Measuring
Real-time determination of position is a fundamental prerequisite for many applications in today’s society, not least in aviation and automotive industry. There are many ways to accomplish the positioning, more or less suitable depending on the application. Lately, the interest in unmanned air vehicles (UAV) has increased rapidly [22]. The position determination of those vehicles are of high importance due to the fact that they are not controlled by an on-board pilot.
4.2.1
Global Positioning System (GPS)
The Global Positioning System (GPS) is a navigation system based on satellites. The satellites was originally put into orbit by the U.S. Department of Defense in 1973, and are today available for both military and civilian use. The satellites are available anywhere in the world, 24 hours per day. GPS works in all types of weather and consists of three segments: Space Segment; Control Segment and User Segment. [7]
4.2.1.1 Space Segment
There are always at least 24 satellites available, placed in six orbital planes, see Figure 13. The inclination relative to the equator is 55◦and there are four satellites in each orbital
plane. This 24-satellite constellation ensures the users ability to view at least four satellites at the time, at any time and any point on Earth. At some optimal time slots, up to 12 satellites can be used at the same time, which increases the accuracy highly. [23] The satellites operates at an altitude of 20 200 km and each satellite circles the Earth twice a day. The number of operational satellites are continuously increased, and as of 9th January 2019, there were 31 satellites in operation. [7]
Figure 13: The 24-satellite constellation in six orbital planes [7]
Each satellite has its unique signal and orbital parameters. By sending those to a GPS receiver, the precise location of the satellite can be determined. The receiver can then by using this information calculate the exact position of the object. Once the position of the object is found, other information such as speed, time and bearing can be calculated. [24] In order to track the two-dimensional position (longitude and latitude) of an object, a GPS receiver on the object must be in contact with the signal of at least three satellites. In cases where the three-dimensional position should be determined (longitude, latitude and altitude), a minimum of four satellites are required. [7]
The GPS originally sends data on two different frequencies: L1 = 1575.42 MHz
L2 = 1227.60 MHz
The civil frequency, L1, is used for both civil and military use, while the L2-signal only is available for military applications. The military uses the civil signal to send P-code (Precision-code) in order to increase the safety and lower the risk of the signal to be intruded. The L1 frequency sends C/A-code (Coarse/Acquisition-code). [23]
4.2.1.2 Control Segment
The control segment of the GPS is the Operational Control System (OCS), which includes a master control station, monitor stations and ground control stations [23]. These elements and there correlation is shown in Figure 14. The main purposes of the OCS is to track the satellites for determining the orbit and clock, synchronize the time of the satellites, and upload the data.
There are in total 16 monitor stations available. Each of these tracks the position of all satellites passing overhead. [7] There are two mathematical ideas behind the GPS positioning network; Trilateration and Pseudorange. [25]
Figure 14: A schematic view of the control segment elements and there correlation [7]
Trilateration calculates the position of an object by measuring the distance from different satellites to the object. The object’s GPS receiver acquires a specific time and distance from each satellite available at the moment. Since the location of all satellites are known, this distance gives information about where the object is located. The distance to each satellite is known, though the angle is still unknown. This means that the object can be located anywhere on the circle with a radius equal to the distance from the satellite, see Figure 15.
Figure 15: Fundamentals of Trilateration[8]
To get the exact two-dimensional position of the object, the same process must be done with two more satellites. Each satellite is at the center of a circle and where they all intersect is the position of the object, see Figure 16. Though, the Earth is three dimensional, which means that the GPS satellites broadcast signals as a sphere, shown in Figure 16. The three-dimensional position of the object can be determined by using four satellites. The point where all four spheres intersect is the position of the object. [25]
Figure 16: Intersection between different satellites [8]
The other mathematical idea, Pseudorange, is the time the signal takes to travel from the satellite to the receiver, multiplied with the speed of light. Equation (5) states the formula for pseudorange Ps, where T is the time of the receiver clock, Ts the time of the
satellite clock, and c is the speed of light (=299 792 458 m/s in vacuum).
Ps= (T − Ts)c (5)
Since the atomic clock on the receiver is not totally reliable, clock errors often occurs, thereby the range definition pseudo. At least four satellites must be used to solve the pseudorange equation and find out the three-dimensional position of the receiver. With four unknowns and four equations the method of least-square is used to determine the object’s position.
4.2.1.3 User Segment
The user segment consists simply of the GPS receiver equipment, which receives the signals from the satellites and uses the transmitted information to calculate the user’s three-dimensional position.
4.2.2
Differential GPS
Differential GPS (DGPS) is an improvement of the original Global Positioning System which provides highly improved position accuracy. DGPS uses fixed ground-based reference stations to adjust the real-time GPS signals in order to eliminate the pseudorange errors. When the satellite signals are travelling through the atmosphere heading to the GPS receiver, they are subjected to delays. Since the pseudorange is based on the travel time of the signal, this results in an incorrect pseudorange which leads to errors in the position measurements. A relative small time delay can lead to a large position error which means that the continous changes in the atmosphere can result in a very inaccurate position depending on the present conditions. [26]
The Differential GPS can either be integrated directly in the GPS receiver used in the object, or it could be added afterwards when post processing the data. In the latter case, the GPS receiver just collects all measured positions and the time for each measurement. This data can then be processed afterwards and merged with the corrections made by the reference station.
4.2.3
Real-Time Kinematic (RTK)
Real-Time Kinematic is a carrier-based ranging method that gives a higher accuracy than the code-based GPS technique. The basic concept of the RTK technique is to reduce and remove the common errors that occurs due to atmospheric changes. [9] By using a carrier wave signal instead of a code-signal as used in GPS, a more precise signal change can be obtained. This enables knowledge about exactly when the signal was changed, which gives a much higher timing resolution. When merging the real-time and corrected signals, the time change is highly accurate which means that a very precise position can be calculated by comparing the distance between satellite and receiver. [27] Figure 17 shows the principal of Real-Time Kinematic.
Figure 17: Principle of real-time Kinematic [9]
4.2.4
Inertial Navigation System (INS)
An Inertial Navigation System (INS) is a device that uses motion sensors, rotational sensors and a computer to measure and calculate the three-dimensional position, velocity and acceleration of an object. Usually, the motion sensors are accelerometers and the rotational sensors are gyroscopes. The accelerometers and gyros provides measures of the object’s acceleration and angular velocity. By receiving those measures, the navigation computer can then calculate the relative position, orientation and velocity of the object. The INS is suitable for a wide range of applications, such as navigation of aircraft, missiles, spacecraft and marines. The largest difference between the GPS and INS is that the GPS generates an absolute position using a known coordinate system, while the INS generates the position relative to the last known point. [28]
By definition, an INS consists of two essential parts; the Inertial Measurement Unit (IMU) and the navigation computer. The IMU comprises the accelerometers and gyros that performs the measurements and serves the navigation computer with information. [28] An important aspect when talking about INS is how the reference frames are defined. In this thesis, the reference frames are defined as shown in Figure 18; the navigation system’s reference frame as the body frame, and the frame of reference in which the navigation is made as the global frame. [10]
The IMUs can be divided into two main categories; Stable Platform Systems and Strap-down Systems, where the only difference is the frame of reference where the accelerometers and gyros are operating.
Stable Platform Systems
In a Stable Platform System, the inertial sensors (accelerometers and gyros) are mounted on a platform that are always held aligned with the global frame of reference. [10] This means that the part where the sensors are attached is isolated from all angular motions created by the object, for example an aircraft. [29] The platform is mounted using gimbals (frames), which allows the platform to be free in all three axes, shown in Figure 19.
Figure 18: The definition of body and global frames in relation to the object [10]
Figure 19: Example of a stable platform IMU [10]
The gyroscopes attached to the stable platform detects eventual rotations and reports those to the torque motors which then rotates the gimbals to maintain the platform aligned with the global frame. [10] Due to many rotating parts and a complex physical shape, the maintenance of a stable platform IMU is expensive and time-consuming [30].
The orientation is tracked by reading the angles between the gimbals using the angle pick-offs. The position of the object can be calculated by integrating the accelerometer signals twice. Before the signals from the accelerometers are integrated, they are corrected for the gravitational impact. The inertial navigation algorithm for a Stable Platform System is presented in Figure 20.
Strap-down Systems
In a Strap-down System, the inertial sensors are mounted rigidly, or "strapped down", to the frame of the object, shown in Figure 21. [30]
Figure 21: Example of a Strap-down IMU [11]
This enables the INS platform to follow the movements of the object which it is attached to. Hence, the sensors are measuring the acceleration and angular velocity in the body frame. A Strap-down System are in general less complex than a Stable Platform System, both regarding physical shape and computation. [10] Through less moving parts, the maintenance and reliability over time is improved. Even though the Strap-down Systems requires more accurate inertial sensors and more computing power, the benefits of lower cost, complexity and weight make these systems the optimal choice within aviation. [30]
Figure 22: The inertial navigation algorithm for a Strap-down System [10]
Figure 22 shows the inertial navigation of a Strap-down System. The physical principles are the same as for the stable platform showed in Figure 20. For the Strap-down System, the orientation is determined by integrating the gyroscope signals. The accelerometer signals are transformed into global coordinates using the orientation obtained from the integrated gyro signals. When the global acceleration signals are obtained, the position can be determined by the same procedure as the Stable Platform IMU. [10]
4.2.4.1 Accelerometer
An accelerometer is an electromechanical device that measures proper acceleration. Proper acceleration is the acceleration of a body relative its own rest frame, meaning the coordinate system in which the body is at rest. The forces measured by an accelerometer can be both static, for example gravitational forces, or dynamic to sense movement and vibrations. [31] An accelerometer in free fall measures an acceleration of 0 m/s2, while one placed on the Earth’s surface will read -9.82 m/s2. In the case of free falling, both the mass and the casing are falling by the action of gravitational forces, which means that the relative difference is zero. When placed on the surface of Earth, the forces acting on the casing is zero while the mass is effected by the gravitation, resulting in a negative acceleration. [28] Normally, an accelerometer consists of some type of capsule with a weight inside. The capsule is firmed into the object which acceleration should be measured, while the weight inside is loose. Due to the inertia of the weight, pressure forces, drag forces and displacements occurs between the weight and the capsule, which then transforms into electrical signals. A simple accelerometer usually consists of a mass attached to a spring, that measures the deflection and converts it into acceleration. Accelerometers can be used in many different applications such as geology, navigational and inertial guidance systems, smartphones and industrial vibration measurement. [32] Figure 23 shows a typical light weight accelerometer.
Figure 23: A typical light weight accelerometer [12]
The position x of an object at any time t can be determined by estimating the acceleration of the object, as stated in Equation (6),
x(t) = t Z 0 t Z 0 a(t)dtdt + t Z 0 v0dt + x0 (6)
where a is the acceleration of the object, v0 is the initial velocity of the object and x0
the object’s initial position. [33] The acceleration a, can by using accelerometers be measured directly. Though, one problem with accelerometers is the large amount of errors that occurs in the measurements. The extent and impact of these errors must be taken into consideration when using this device for applications where the accuracy is a critical parameter. The direct output from an accelerometer is an electrical signal proportional to the acceleration of the device, often measured in voltage. [33] A simplified model of the measured voltage is
Vm(t) = S(1 + δS)a(t) + Vb+ δVb (7)
S is the sensitivity δS is the sensitivity error Vb is the voltage bias
δVb is the voltage bias error.
By rearranging Equation (7) to solve for the measured acceleration, Equation (8) is obtained. [33] am(t) = Vm(t) − Vb S = a(t) + δSa(t) + δVb S (8)
The initial velocity and position (v0 and x0) are assumed to be zero at t = 0. Combining Equation (6) and (8), and integrating twice gives the time dependent position of the object as x(t) = t Z 0 t Z 0 am(t)dtdt = t Z 0 t Z 0 a(t)dtdt + δS t Z 0 t Z 0 a(t)dtdt +δVb S t Z 0 t Z 0 dtdt. (9)
The first term of Equation (8) is the data desired to obtain the position of the object. The second and third terms are the internal errors that occurs in the sensor. It can be seen from these terms that the errors are time dependent and increases quadratically with time. This results in a rapidly growing error over time. For example, a position error of 0.1225 meters after one second, has growth to an error of 3 meters in five seconds. An accelerometer with a three-axis sensor can measure the position in all three dimensions (latitude, longitude and altitude). [33]
In general, an accelerometer by itself is not a good option for position measuring of an aircraft, due to the large gravitational variations. Those variations gives very unstable data and the accuracy becomes low. Though, using accelerometers in combination with gyroscopes and a computer (INS) is usually a good choice for aircraft applications, due to the high accuracy obtained.
4.2.4.2 Gyroscope
A gyroscope is a device which uses gravity to measure and determine orientation of an object. There are a lot of different types of gyroscopes, for example mechanical, optical and vibrating. The most conventional one is the Mechanical gyroscope, which consists of a spinning wheel mounted on two gimbals, showed in Figure 24. The gimbals allows the wheel to rotate in all three axes. The Conservation of Angular Momentum principle, allows the wheel to remain stationary and resist changes in orientation. By using angle pick-offs, showed in Figure 19, the angles between the gimbals can be measured, and the orientation of the object can be identified. A large disadvantage with mechanical gyroscopes is that the moving parts causes friction, which results in a higher maintenance cost and less accurate output results over time. [13]
Unlike the mechanical one, the Optical gyroscopes and MEMS (Micro-Electro-Mechanical gyroscopes) measures angular velocity. A fibre optic gyroscope consists of a large coil of optical fibre, and measures the angular velocity by using the impact of light. The basic principle is based on the Sagnac Effect. The theory describes the physical effects that occurs when two light beams are fired into the coil in opposite directions. If the coil is stationary, both beams will travel the same path in the same time, independent on the direction. Though, if the coil is under rotation, the direction of travel of the beams will
Figure 24: Example of a mechanical gyroscope [13]
govern the path and travel time around the coil. The beam travelling against the direction of rotation will have the shortest path and travel time, showed in Figure 25. [34] When the beams reaches the exit of the coil, they are combined into one, due to the phase shift created by the Sagnac Effect. By measuring the intensity of the combined beam, it is possible to determine the angular velocity. [10] In comparison with the mechanical gyros, the optical ones are easier to maintain due to no moving parts. The accuracy of an optical gyroscope is dependent on the size, larger ones has higher accuracy. [10]
Figure 25: Visualization of the Sagnac Effect. The dashed line shows the path taken by the beam travelling with the direction of rotation [10]
A Ring Laser Gyro (RLG) uses the Sagnac Effect as well, with the difference that instead of optical fibre, mirrors are used. In modern aircraft the RLG is the predominant gyroscope type used. [30]
Another gyroscope type is the MEMS, which is a combination of mechanics and electronics. A MEMS is using vibrating elements to measure the Coriolis effect, which is the principle force that acts on a mass moving with a velocity v in a frame of reference that is rotating with an angular velocity ω. [10] The simplest variant of a MEMS gyroscope consists of two masses oscillating and constantly moving in opposite directions, showed in Figure 26. By applying an angular velocity, the Coriolis force on each mass will act in opposite the direction. This will result in capacitance change, that is proportional to the angular velocity. [14]
Figure 26: Coriolis effect when angular velocity is applied [14]
There are many advantages with these gyroscopes, such as the small size, low weight, few parts and low power consumption. Despite that, they are not suitable for applications where high accuracy is required, for example in aircraft. [10]
4.2.5
Combined GPS/INS with Kalman Filter
In reality, both the INS and GPS data comes with inaccuracies. The GPS is highly dependent on location and time, since the number of available satellites varies through the day. These outages can be reduced by the use of an Inertial Navigation System, since the INS are using dead computation and are totally independent on external sources. The INS on the other hand, often contains low-cost inertial sensors that exhibits large errors. Those errors can be compensated using position and velocity data from a GPS. The combination of GPS and INS, with the use of a Kalman filter, results in a high-accuracy real-time navigation system. [35]
A Kalman filter is a set of mathematical algorithms, that based on a multitude of incomplete or uncertain measurements can measure the state of a dynamic system. Kalman filter can be used for many different applications, but are most valuable for systems which are continuously changing. One example of such system is an INS used to continuously measure the position and velocity of a moving object. The Kalman filter can then produce correct and accurate information about the object’s position and velocity, based on a series of imperfect observations made by the inertial sensors. The filter can estimate the current optimal state of a system, by combining the measured output with a known model of the dynamic system. [36]
The position and velocity measurements from the GPS or inertial sensors, comes as a range of possible positions and velocities for each time stamp. Some of the data points might be true, and it is the Kalman filter’s task to find out which position and velocity that is real. The filter assumes that the positions and velocities are random and distributed by normal distribution. The basic representation of the Kalman filter is a linear estimator, and is only applicable on linear systems. In order to filter the data for navigation systems as INS or GPS, the Extended Kalman filter must be used, which linearizes about an estimate of the current mean or co-variance of the variables. [15]
Define the state x ∈ Rn of the object as a vector of the position and velocity. Due to the non-linearity, the process is governed by the non-linear stochastic difference equation
xk+1= f (xk, uk, wk) (10)
with a measurement z ∈ Rm that is
zk= h(xk, vk). (11)
In Equation (10) and (11), the random, normal distributed variables wkand vkrepresent the process and measurement noises. The non-linear function f (•) in the difference equation (10) relates the time at step k to the state at step k + 1. The non-linear function h(•) in the measurement equation (11) relates the state xk to the measurement zk. [15]
Since the values of the noise wk and vk is not known at each time step, the state and
measurement equations can be approximated without them as ˜
xk+1= f ( ˆxk, uk, 0) (12)
and
˜
zk= h( ˆxk, 0) (13)
In order to linearize an estimate about Equation (12) and (13), new governing equations are defined: [15]
xk+1≈ ˜xk+1+ Ak(xk− ˆxk) + Wkwk (14)
zk≈ ˜zk+ Hk(xk− ˆxk) + Vkvk. (15)
The new variables used in Equation (14) and (15) are described in Table 1. [15]
Table 1: Description of variables used in Kalman equations
Variable Description
xk+1 and zk actual state and measurement vectors
˜
xk+1 and ˜zk approximate state and measurement vectors
ˆ
xk a posteriori estimate of the state, from a previous step k)
Ak the Jacobian matrix of partial derivatives of f with respect to x
Wk the Jacobian matrix of partial derivatives of f with respect to w
Hk the Jacobian matrix of partial derivatives of f with respect to x
Vk the Jacobian matrix of partial derivatives of f with respect to v
The predicted error is defined as ˜
exk ≡ xk− ˜xk (16)
and the measurement residual as ˜
ezk ≡ zk− ˜zk. (17)
By the use of Equation (16) and (17), the governing equation for the error process can be stated as ˜ exk+1 ≈ Ak(xk− ˆxk) + k (18) and ˜ ezk ≈ Hke˜xk+ ηk (19)
where k and ηk are new random variables with zero mean, presented below together with the approximated probability distribution for the error prediction ˜exk
p(˜exk) ∼ N (0, E([˜exke˜ T xk])
p(k) ∼ N (0, W QkWT)
p(ηk) ∼ N (0, V RkVT)
Combining Equation (16), (17) and (19) results in a posteriori state estimate for the non-linear process as
ˆ
xk= ˜xk+ ˆek. (20)
Given the approximated probability distributions and letting the predicted value of ˆek
to be zero, gives
ˆ
ek= Kkeˆzk (21)
where Kk is the Kalman gain, that minimizes the a posteriori error covariance defined as Pk= E[ekekT]. On of the forms that minimizes the error covariance is stated in Equation
(22). For more details about the derivation of the gain can be found in [37].
Kk= ¯PkHkT(HkP¯kHkT + VkRkVkT)−1 (22)
Substituting Equation (21) into (20) with the use of Equation (17) results in ˆ
xk= ˜xk+ Kk(zk− ˜zk). (23)
All equations derived and stated in this section, finally leads to the time update and measurement update equations for the Extended Kalman Filter. [15] The final equations are summarized in Figure 27.
Figure 27: Summation of the time and measurement update equations used in the Extended Kalman Filter[15]
4.3
Data Acquisition
The aim of the system is to measure and collect information about the position and velocity of the store, from separation to impact. The information should both be stored and transmitted to a ground station in real-time.
4.3.1
Computer storage
A posteriori analysing and verification of results are often desired in flight testing. Therefore, the data measured during the test must be collected and stored in some way. The most suitable way to accomplish this is to use a data logger or a SD card.
A data logger is an electric device that records data over time or relative a location, either with built in sensors or via external sensors. The logger is usually a small, battery powered digital processor with an internal memory. One benefit with using a data logger is the ability to collect data during a long time upon activation. [38]
Secure Digital (SD) is another way to collect and store the data acquired by the sensors in the data acquisition system. A SD-card is a non-volatile type of computer memory often used in portable devices. [39] The most significant difference compared with the data logger is that a SD-card does not require constant power in order to retrieve and store data. Therefore, it is often the most suitable type of memory in applications where the battery capacity is limited. [40]
4.3.2
Data Transmission
Transmitting data from an aircraft to a ground station enables real-time analysis and verification of the present system, which is often valuable in flight testing. Within the aircraft industry, telemetry is often used as a general description for transmission of data. As well in the health technology sector, telemetry is a valuable tool used for example in continuous heart rate monitoring.
Telemetry is an automated communication process where measured data are wireless transmitted from the object of measurement to a receiver station. For space and aircraft applications, long distance telemetry systems are used. In long distance telemetry, used for example in flight testing, data from the different sources are sampled and sent over a unified channel.
Figure 28: Schematic picture of a telemetry system [16]
A telemetry system consists of a transmitter and a receiver, with corresponding antennas. The transmitter collects the measured data from the sensors in the system, unifies them into one data package and sends them to a telemetry receiver unit. The receiver acquires the measurement data and outputs a filtered data stream. [16] Dependent on the application and the required accuracy of measurements, the data could be further processed afterwards, for example with Differential GPS as mentioned before.
4.3.2.1 Friis’ Transmission Equation
There are a large amount of different transmitters, receivers and antennas on the market. Which components to use are highly dependent on the specific application. In the case of a flight test, the maximum range between transmitter and receiver is often specified since the altitude of flight is known. In other cases, the knowledge about the length of the transmission range is desired, in order to specify where to place the receiver antennas for a specified system. Independent on which parameter that is unknown, the relation between transmitter power, receiver power and range is the same. Equation (24) is called Friis’ Transmission Equation [41], PR= PTGTGR λ 4πR 2 (24) where
PR is the receiver power [W]
PT is the output power from the transmitter [W]
GR is the receiver antenna gain [-]
GT is the transmitter antenna gain [-]
R is the range between transmitter and receiver antennas [m] λ = fc is the wavelength of the signals [m].
The equation states the relation between the power received from one antenna when transmitted from another antenna. [41] Each antenna has a gain, which is simply a performance parameter that describes how much power that is needed to convert radio waves into electrical power (receiving antenna) or convert input power to radio waves (transmitting antenna). [42]
By examine Equation (24), it is obvious that the receiver power is dependent on all other parameters in the data transmission system. The choice of antennas is highly dependent on the actual application of use. There are two main types of antennas; Omnidirectional and Directional. An Omnidirectional antenna radiates and receives signals equally well in all horizontal directions, while a Directional one is able to focus the signal reception in a particular direction. [43] A high gain means a high directivity, in other words how directional the antenna’s radiation patterns is. [44] Therefore, an antenna with high gain is suitable in cases where the desired signal direction is exactly known. For applications where there are several receiving units or moving objects, a low gain antenna is better to use since the radiation pattern is more spread.
Figure 29 shows two different types of antennas, one Omnidirectional (dipole) and one Directional (yagi), with the corresponding signal patterns in the horizontal direction. 4.3.2.2 Frequency Range
When sending data between two objects, the signals are sent over a specific frequency, which basically represent the number of cycles per unit time. Depending on application, either a high or low frequency is most suitable. The International Telecommunication Union (ITU) has divided all available frequencies into bands with different frequency ranges. [45] As stated in Figure 30, there are nine frequency bands [45], although only two of them are
Figure 29: Comparison of antenna types
appraised to be suitable for the acquisition system developed in this thesis. This is based on the required distance of communication along with the volume of data to be sent.
Figure 30: The frequency bands defined by ITU. The range applicable in this thesis is circled with green
In general, a lower frequency results in a longer range due to its ability to go around obstructions in the atmosphere. Lower frequency radio waves exhibit low signal attenuation, which means that the gradual loss of flow intensity through a medium is low. A higher frequency requires the receiving and transmitting objects to be within line of sight with each other, which means a lower range. However, more data can be sent on a higher frequency since the update rate is higher.
5
Methodology
Based on the concept development processes presented in Section 3, an adapted methodology was produced for this thesis. The methodology, presented in Figure 31, includes parts from the different methods developed by Liedholm, Ulrich & Eppinger, and Ullman. It was divided into four general phases; Planning, Function Analysis, Concept Generation and Concept Evaluation.
Figure 31: Methodology adapted for this thesis
The methodology adapted for this thesis was mostly based on the concept development process according to Ulf Liedholm [6]. However, the Concept Generation phase evolved by Liedholm was here divided into two comprehensive phases in order to separate the problem definition and the concept generation. The Function Analysis was made fully according to Liedholm’s method, where the problem was decomposed into functions and sub-functions by creating a Black Box and a Function/Mean-tree.
The Concept Generation phase was a combination of Liedholm’s and Ulrich & Eppinger’s methods. As Liedholm states, the concepts should be established by combining the means defined in the Function/Mean-tree [6]. Ulrich and Eppinger suggests a less structured way where the developer should search for solutions by interviewing lead users and search for patents [4], which has been the most useful idea-generator in this project. To evaluate the concepts a requirement check was made according to Liedholm, and then a list of advantages and disadvantages was produced.
