Researchers from Massachusetts Institute of Technology upgraded the Cheetah 3 robot to operate blindfolded.
Cheetah 3- a robot developed by MIT is upgraded to operate without relying on cameras or any external environmental sensors. This revolutionary ability is credited to two new algorithms developed by the team: a contact detection algorithm, and a model-predictive control algorithm. It enables the robot to make its way up staircases and walk through unstructured terrain blindly. It can quickly recover its balance due to unexpected forces with assistance from the contact detection algorithm. This algorithm determines the best time for a given leg to switch from swinging to stepping on the ground. It is a probability based algorithm that constantly calculates for each leg three probabilities that determine whether the leg will make contact with the ground, the force required once the leg hits the ground, and the probability that the leg will be in mid swing. The algorithm was tested with the Cheetah 3 trotting on a laboratory treadmill and climbing on a staircase. To deduce strong results both surfaces were littered with random objects such as wooden blocks and rolls of tape. The robot will be presented at the International Conference on Intelligent Robots 2018 at Madrid in October.
Cheetah 3’s blind locomotion is also a result of the model-predictive control algorithm that predicts the amount of force required to step without losing balance. This algorithm calculates the multiplicative positions of the robot’s body and legs to determine if a certain force is applied by any given leg as it makes contact with the ground. The calculations for each leg are made with a frequency of 20 times per second. To test the algorithm, unexpected forces were introduced by kicking and shoving the robot while it trotted on a treadmill, and yanking it by the leash while it walked up the staircases with obstacles. The robot was quick to produce counter-forces to regain its balance and kept moving forward.