Projects

1
Projects

• EECS290p

2
Requirements

• Projects must be related to one of the following
  Topics
• Control,
• Sensing, or
• Algorithms
• Projects must have a major
• Software component,
• Theory component, or
• Testing component
• Individuals will
• Perform work on their own, or in small groups
  (e.g., 2)
• Join together under an umbrella topic (e.g.,
  control) to produce a final presentation
• Documentation in the form of
• Project proposal (describing scope)
• Project design document
• Implementation document (final contribution)
• Deliverables as determined by Project Topic and
  Instructor

3
Dates

• 15 September Project Teams
• List of team member(s) with umbrella topic (e.g.,
  control)
• 29 September Project Proposals
• Member(s), Topic, Scope, Deliverables, Timeline
• 9-13 October Meetings w/ Instructor
• 27 October Project Detailed Designs
• Member(s), Topic, Scope, Deliverables, Timeline,
  Progress, Implementation Details, Topic-specific
  material (simulations, s/w architecture, early
  results)
• 4-8 December Project Presentations
• 8 December Delivery of Final Project
• Member(s), Topic, Scope, Deliverables

4
Possible Topics Control

• Control layers design/implementation.
• hardware layout,
• steps for verification of simulations,
• design for hwil, swil testing
• Controller software modules
• module design for layers of control
• required software modules for actuation through
  off-the-shelf hardware
• model-predictive control in layers
• ports of controller modules to Orca2
• State data rates and requirements
• analysis of requirements for performance at
  certain speeds
• use of control software modules
• proofs of performance and schedulability
• Tests and diagnostics
• System identification techniques
• Techniques to verify plant models
• Experiments to determine or analyze robustness

5
Possible Topics Sensing

• Data processing algorithm requirements
• development of lists of requirements
• design of dataflow for sensor throughput
• analysis of sensor layout, and coverage
• Software modules for sensor fusion
• module requirements
• testing and diagnostics of fusion baselines
• Localization technique verification
• simulation techniques for error localization
  tests
• RDNF/MDF localization coordination
• strategies determining truth from conflicting
  sensor data
• Regression testing failures
• bootstrapping techniques to debug failures
• test setup using existing software infrastructure

6
Possible Topics Algorithms

• Computer vision
• layers for distributed computation
• FPGA implementation of existing algorithms
• AI (various)
• planning
• optimization
• Code generation
• safe set analysis and runtime support
• Graphics/Rendering
• rapid worldview update
• representation isolation from display

7
Quick Demonstration

• Orca2
• Software Infrastructure
• http//orca-robotics.sourceforge.net/