Collision avoidance systems are used in many applications like robots, electric wheelchairs, and autonomous vehicles for navigating around static obstacles. Dynamic window approach (DWA) is popular and known as a collision-free navigation algorithm for static obstacles, but it will often collide with moving obstacles with inaccurate trajectory predictions. Existing DWA needs trajectory predictions for collision avoidance for moving obstacles, such as using 2 seconds past information to predict future 3 seconds trajectory. However, existing, and current technologies are limited to avoiding collisions with dynamic objects.
Researchers at Rutgers have developed a novel algorithm that enables autonomous devices to actively avoid moving obstacles. Prototypes using the system demonstrated the ability to efficiently determine admissible velocity of the agent considering the mobility of moving obstacles. The invention removes the trajectory prediction and guarantees collision-free for moving obstacles.
- Autonomous vehicles
- Electric wheelchairs
- Avoid collisions with moving obstacles without the need of trajectory predictions.
- Improves the safety of cars, robots, and patient mobility devices.
Intellectual Property & Development Status: Patent pending. Available for licensing and/or research collaboration. Please contact email@example.com
TTO Home Page: https://techfinder.rutgers.edu
Name: Andrea Dick
Title: Associate Director, Licensing
Department: Innovation Ventures