Collisions occur in nature all the time. They range from cellular collisions in cell migration to humans colliding while navigating through crowds. In robotics research, we learned that collisions must be avoided to protect the robot and its surroundings. However, I came to the conclusion that conventional robots avoid collisions because of the way they were designed. We have chosen materials from a list of materials and environments from a list of environments that made robots fast and reliable, but their collisions also destructive. But what happens if we choose differently from the lists of materials and environments?
I started answering this question at the Swarms: From Biology to Robotics and Back workshop at this year’s International Conference on Robotics and Automation (ICRA 2018) in Brisbane, Australia. A 3-minute video presentation [1] in addition to a poster and extended abstract [2] focused on the role of collisions in (swarm) robotics. I presented two examples of unconventional robots that embrace physical interactions and collisions: i) soft robots ii) robots that use active low-friction locomotion such as HoverBots.
If you are interested in studying “collision robotics”, please have a look at my workshop contributions [1] [2]. If you are interested in building your own robotic research tool to study collision dependent robot behaviours, I recommend reading [3] to learn more about how to build robots that embrace collisions, and [4] to learn more about how you make robots detect collisions.
[1] HoverBots: Embracing and Detecting Collisions Using Robots Designed for Manufacturability – workshop video
[2] HoverBots: Embracing and Detecting Collisions Using Robots Designed for Manufacturability – workshop extended abstract
[3] HoverBots: Precise Locomotion Using Robots That Are Designed for Manufacturability
[4] Multi-Functional Sensing for Swarm Robots Using Time Sequence Classification: HoverBot, an Example
Recent Comments