Blind Robotic Beast Prevents Balance Loss Using New Algorithms


Not the actual Blind robotic beast. Photo used only for representation / Photo by: Ociacia via Shutterstock


Engineers at the Massachusetts Institute of Technology built a mechanical beast capable of walking and galloping through various terrain without any kind of vision.

The robotic beast has a weight of about 90 pounds and a body size of a full-grown Labrador, equipped with blind locomotion. The engineers included two new algorithms to allow blind locomotion to work on the robot.

The first one is called the Contact Detection algorithm. This enables the robot to know when is the best time to shift from swinging in the air to stepping on the ground. For instance, if the robot steps on a light branch, it continues to step on the ground, but if it makes contact a hard, heavy rock, it shifts to swinging motion so it can prevent the loss of balance.

The second algorithm is called Model-Predictive Control. This algorithm aids the robot to best determine the transition of one leg between step and swing motions. It consistently calculates every leg of three probabilities – the probability of making contact with the ground, the probability of the force created when a leg hits the ground, and the probability that a leg will swing in mid-air. All of these predictions are based on data collected by several accelerometers, gyroscopes, and joint positions of the legs.

“It doesn’t know the height of each step, and doesn’t know there are obstacles on the stairs, but it just plows through without losing its balance. Without that algorithm, the robot was very unstable and fell easily,” explained Sangbae Kim, the designer of the robot and an associate professor of mechanical engineering at MIT.

Each algorithm measures data on each leg for every 50 milliseconds. When they tested the robot by kicking and shoving it as it stepped on the treadmill, the algorithm helped the robot to perform counter-forces to prevent balance loss and sustain its movement.