by Eric Watson If you want to borrow some code, go for it. If you want to contribute, fork away and send me a PR or email me at watson.eric.r@gmail.com to discuss direction.

Improve my hardware skills, learn more python, and sharpen simulation skills while having some fun and learning about autonomous vehicle control, swarm intelligence, and inter-vehicle communication. Goals

Stage 1: One autonomous car completes navigation objective Stage 2: Add a second autonomous car without sensors in following formation with the first using inter-vehicle communication Stage 3: Add sensors and intelligence to both vehicles to enable efficient swarm formation driving

Starting with one stage at a time: simulate dynamics and control, then move to processor-in-the-loop (pyboard), and then full hardware runs with vehicles. Stick to this as closely as possible considering I’m living out of a backpack at the moment!

One autonomous car completes navigation objective

Car A (Alfred):

• Sensors and smarts to complete navigation goal → find brightest light
• Pyboard, light sensors, battery and chassis required Simulation
• Python simulation including:
• Light source static location
• Sensor readings (left or right is brighter?)
• Simple vehicle dynamics model (constant velocity, max turn rate)
• Bang-bang control turning the vehicle toward brighter light
• Store and visualize light source location, sensor readings, vehicle location, and control values Then, bring processor in the loop. Then, full hardware!

Add a second autonomous car without sensors in following formation with the first using inter-vehicle communication

Car A (Alfred):

• Sensors and smarts to complete navigation goal → find brightest light
• Add transmitter to share state with Bert Car B (Bert):
• “Knows” Alfred’s state and follows in formation
• Pyboard, battery, receiver, and chassis required

Simulation

• Sharing of control inputs from Alfred to Bert
• Similar Bert vehicle dynamics model with follow control implemented
• Store and visualize light source location, sensor readings, vehicle locations, and control values Then, I’ll bring both processors in the loop with communication between them. Then, full hardware!

Add sensors and intelligence to both vehicles to enable efficient swarm formation driving

All Vehicles (Alfred & Bert)

• Add GPS receivers and 2-way communication sharing state with each other
• Swarm intelligence rules to formation drive to navigation objective

Simulation

• Share states between vehicles
• Implement formation driving rules
• Store and visualize light source location, sensor readings, vehicle locations, control values, and relative dynamics Then, I’ll bring both processors in the loop with communication between them including simulated GPS data. Then, full hardware out in a parking lot!