New Castle Design Associates Concept Review of Insect Video Tracking Device October 27, 1998 Team 5 Sponsors: Keith Hopper, USDA & UD

Project Information Members Primary Customer Justin L. Combs Keith Hopper (USDA and UD) Raymond M. Foulk IV Ryan S. McDonough Advisor George H. Sapna III Dr. James Glancey

Summary Mission: Our mission is to design, construct, and refine an insect video tracking system for agricultural research that provides our customers with a creative, realistic and performance-based solution. Approach: Our strategy will be to gain an overall knowledge of the project and then to strive for a solution by researching, benchmarking, and defining the customer’s wants and constraints. Finally, using an iterative design synthesis process, our team will generate the best solution to satisfy our customers.

Background $100 million crop damage each year due to pests Introduction of beneficial (predatory) insects into environment Study of reproductive habits - Aphelinus asychis

Problem Description Existing System Measures insect movements within a small arena Camera Specimens Computer Problem With Existing System Disturbs the behavior of the insects

CustomersWants Keith Hopperlarge area, position/speed, not disturb insect, track for 20 min., existing equipment, C++, minimize post pro., wireless device, interchange camera Mike Smith * position/speed, minimize post pro, various conditions, reduce pesticides Richard Turcotteadaptable to other insects, not disturb insect, simple interface, large area, position/speed * Customer added recently.

CustomersWants Mr. Filaskylow priced produce, reduce pesticides, unharmful insects Rex Mearshigh % corn yield, reduce pesticides, unharmful insects Alice Klinelow priced produce, improve sales, reduce pesticides Greg Frantzreduce pesticides, benefits outweigh costs, adaptable to other insects Anthony Wexlerreduce pesticides, energy efficient, unharmful insects

Previous Top 10 Wants Track the insect over large area Measure position and speed of the insect Adaptability to other types of insects Not to disturb the insect’s behavior Easy-to-learn user interface Ability to record for minutes Maximize use of existing equipment Benefits must out-weight costs Preferred language is C/C++ Minimize post-processing of data

Revised Top 10 Wants Measure position and speed of the insect Track the insect over large area Adaptability to other types of insects Minimize post-processing of data Not to disturb the insect’s behavior Easy-to-learn user interface Ability to record for minutes Use existing equipment Benefits must outweigh costs Preferred language is C/C++

Constraints Previous constraints System must cover 1 square meter Two dimensional tracking system Revised Constraints Project must be completed by end of school year Project expenses must remain below $5000 Must cover a larger area than existing system Work area must occupy only Stearns Lab

System Benchmarking Existing Video Tracking System Ultrasound Scanner at the C.C.M. NASA Vertical-Spin Tunnel Autonomous Robot “RHINO” Semi-Automated Film/Video System

Position/speed X,Y coordinates of insects Instantaneous error Accumulation of error Maximum speed of motion Large areaSize of tracking area Adaptability to other insectsInsect size range Maximum speed of motion Minimize post-processingTotal acquisition time Frequency of acquisition Feedback delay Do not disturb insectDistance from device to insect Smoothness of surface Variation in luminance Wants Metrics

Simple user interface Desired programming language graphical user interface Record for min.Total acquisition time Frequency of acquisition Feedback delay Use existing equipmentAmount of existing equipment used Benefits outweigh costs Savings/Costs C/C++ Programming language Wants Metrics

Functional Benchmarking Torch location algorithm with ATP project at C.C.M. Stepper motors Servo motors with position sensors P.I.D. control Fuzzy Logic control

Old Top 10 Metrics & Target Values 1. Size of Tracking Area1m x 1m 2. Total Acquisition Time20 min 3. Adaptability to Other Insects1mm - 4cm (2D) 4. Frequency of Acquisition1 Hz 5. Feedback Delay0.5 sec 6. Desired Programming LanguageC or C++ 7. Accuracy of Positional Measurements+/- 1mm 8. Distance from Device to Insect0.5 m 9. Amount of Existing Equipment Used100% 10. X, Y Coordinates of InsectYes

Revised Top 10 Metrics & Target Values 1. Size of Tracking Area (7.86%) 1m x 1m 2. Distance from Device to Insect (7.38%) 0.5 m 3. X, Y Coordinates of Insect (6.90%) Yes 4. Savings/Costs (6.43%) 1 5. Maximum Speed of Motion (5.95%) 15 mm/sec 6. Frequency of Acquisition (5.95%) 1 Hz 7. Feedback Delay (5.95%) 0.5 sec 8. Variation in Luminance (5.71%) <5% 9. Total Acquisition Time (5.48%) 20 min 10. Programming Language (5.24%) C or C++ 3a. Accumulation of Error (4.29%) +/- 1mm 3b. Instantaneous Error (4.05%) +/- 1mm

Concept Generation x y z F(s) TF(s) H(s) R C Cartesian Track Polar Track Sensing Surface Robot Wide Angle Pivot Moving Surface Bubble

Concept Evaluation Sorted Metrics Concepts 2. The “best concept” has the highest column total. 3. Repeat step 1 with current “best concept”. Comparison Values Total 1. Compare each concept to the best benchmark.

Concept Selection Previous best concept: Mobile robot with relative positioning Problem: Possible slipping of wheels

Concept Development Potential Solution: Mobile robot with laser positioning Problem: Laser Accuracy Per Cost

Concept Development Preliminary budget for mobile robot with laser positioning:

Relative Concept Quality

Concept Selection Current best concept: Cartesian Tracking System

Development of Cartesian Tracking System Quote for prefabricated system from IDC: $13,100 !

Development of Cartesian Tracking System Configuration: Two Trolleys, Rack and Pinion, & Stepper Motors

Development of Cartesian Tracking System - X Trolley

Development of Cartesian Tracking System - Parts List Estimated Cost: $3, Estimated Shop Time: 120 hours Most Expensive Parts: Rack and Pinion Sets Motors and Controllers Stock Aluminum Other Notes: Included Existing Parts Unaware of Shipping Costs Unaware of Small Hardware Costs

Main Computer Image Analysis Motion Control Algorithm Stepper Controller Insect Position Calculator, Display, & Recorder Digital Camera X-Motor Y-Motor Motor Positions Integrated System Components

Working Models of Cartesian Tracking System

Schedule Highlights Order all partsby Dec. 11 Set up CNC for X-Y trolley end platesby Jan. 08 Build and assemble baseby Jan. 28 Build and assemble X-Y trolleysby Jan. 28 Assemble sub-componentsby Feb. 12 Adapt algorithm to fit physical system by Feb. 26 Test complete systemby Mar. 19 Make changes to the systemby Apr. 14

Schedule To Date

Preliminary Concept 1 Mobile robot with relative positioning: A robot which follows insect over a specified area while recording the insect’s position

Preliminary Concept 2 Stationary high-resolution camera with wide angle lens: A camera records the insect’s motion over a large area

Division of Work Justin CombsConcept Generation & Evaluation Ray Foulk SSD, Benchmarking Ryan McDonoughSchedule, Customer Relations George Sapna SSD, Customer Relations Team EffortPresentations, Planning, Goals Budget $5000 Thousand Dollars Flexible depending upon primary customer’s judgement.

Plans Until Next Presentation 10/ /8 Refine Concepts 11/1-11/15Engineering Analysis of “Best Concept” 11/8 - 11/20Engineering Drawings 11/ /24Written Report 11/ /3Prepare Presentation

Preliminary Concept 3 Cartesian Tracking System Camera moves over a plane in two orthogonal directions. CAMERA INSECT ARENA X Y

Previous Benchmarks Existing Video Tracking System NASA Vertical Spin Tunnel Biorobotic Vision Group DARPA Image Understanding Program Semi-Automated Film/Video System CAD Plotter PC Mouse

Previous Top 10 Wants Track the insect over large area Measure position and speed of the insect Adaptability to other types of insects Not to disturb the insect’s behavior Easy-to-learn user interface Ability to record for minutes Maximize use of existing equipment Benefits must out-weight costs Preferred language is C/C++ Minimize post-processing of data

Design Process BenchmarkingRevise Wants Interview Customers Derive Metrics Concept Generation Concept EvaluationTarget Values Good Solution?

Revision of Wants (Performance Goal) x (Importance to Customer) = (Overall Weight) (Customer Rank) x (Want Factor) = (Importance to Customer) Customers Rank High Want Low Want Exponential Decay of Importance

Derivation of Metrics List of Metrics Top 10 Wants 2. Sum the correlation factors for each metric. 3. Sort the metrics according to total. Correlation Matrix Total 1. Rate the correlation of each metric to each want.

Important Customers Keith HopperEntomologist Mike Smith * Entomologist Richard TurcotteEntomologist Mr. FilaskyFarmer Rex MearsFarmer Alice KlineSupermarket Manager Greg FrantzConsumer Anthony WexlerEnvironmentalist * Added Recently

Comparative Benchmarking 1. Rate each benchmark on how well it satisfies each want. 2. Total the rates to achieve a total score for each benchmark. 3. Sort the benchmarks by total score.