Unlocking the Secrets of Object Permanence: How Engineers are Using Psychology, Physics, and Geometry to Make Robots More Intelligent

Engineers use geometry, psychology, and physics to make robots intelligent
We are surrounded by robots, whether they’re filming video in the air or serving food at restaurants. They can also diffuse bombs during emergencies. Robots improve the quality of life for humans by enhancing their abilities, saving time and improving personal safety and wellbeing. Existing robots have become more adept at simple tasks. However, to handle more complex requests they will need more mobility and intelligence.

Computer scientists at Toyota Research Institute and Columbia Engineering are studying psychology, physics and geometry in order to develop algorithms that will allow robots to adapt to their environment and learn to function independently. This work is crucial to enable robots to tackle new challenges arising from an ageing society and to provide better support for people with disabilities and seniors.

Computer vision has been challenged by object permanence. This is a concept that is well known in psychology. It involves the understanding that an object’s existence is independent of whether or not it is visible. Robots must be able to comprehend our dynamic, ever-changing world. Most computer vision applications ignore occlusions and lose track of objects temporarily hidden.