DJI, uAvionix and Stanford University are working together on sense-and-avoid technologies for unmanned aerial vehicles (UAVs).
Eric Mueller, who works at the NASA Ames Research Center and is a Ph.D. candidate at Stanford University, co-authored a paper with his professor, Mykel Kochenderfer – “Multi-Rotor Aircraft Collision Avoidance Using Partially Observable Markov Decision Processes.”
The paper describes how speed changes by agile multi-copters and small UAVs can be used in addition to horizontal and vertical maneuvers to maintain safe operating distances between aircraft.
The Markov Decision Process (MDP) uses algorithms to model decision-making, such as how an aircraft decides to maneuver to avoid oncoming traffic. MDP is applied in situations where some variables are possibly random (such as how another aircraft may decide to maneuver) and others are under the control of the decision-maker (such as an aircraft’s autopilot).
uAvionix supplied its miniature “ping” automatic dependent surveillance – broadcast hardware for tests at Stanford to determine the effectiveness of the sense-and-avoid algorithms under development.
In addition, DJI supplied the drones for the tests.
“Ping was the only system that provided a realistic way for us to validate our sense-and-avoid algorithms with live traffic for commercial aircraft separation assurance,” says Mueller.
uAvionix recently introduced “pingRX,” a tiny ADS-B receiver that measures less than a square inch. The expected price for pingRX will be $199.