MIT Researchers working on robotic Cheetah have now managed to make it see and jump over obstacles as it runs, just like a human runner would. It was for the first time that any four-legged robot managed to jump over an obstacle, autonomously.

Just last year, MIT researchers showed how well their Cheetah robot can run without any vision technology. Now that the robot has its vision, on board, researchers were able to demonstrate few smart moves performed by this robotic Cheetah, dubbed as “Cheetah 2”.

The robot uses LIDAR technology to sense its obstacles. LIDAR is a visual system that uses reflections from a laser to map the terrain.

The researchers demonstrated the smart moves of the robot in both the indoor as well as an in an outdoor experiment.

In the indoor experiment, the robot was made to run on a treadmill with approaching obstacles, it successfully jumped over the obstacles as tall as 18 inches (that’s more than half of the robot’s height) while maintaining an average running speed of 5 miles per hour.

“A running jump is a truly dynamic behaviour,” says Sangbae Kim, an assistant professor of mechanical engineering at MIT. “You have to manage balance and energy, and be able to handle impact after landing. Our robot is specifically designed for those highly dynamic behaviours.”

Researchers have created a new algorithm to jump autonomously over the obstacles and it is divided into three components.

The first element of the algorithm deals with detecting the obstacle and calculating its height and distance. For that, researchers have developed a formula to simplify environmental details, like representing a path as a straight line and any obstacle as a deviation from that line.

As soon as the robot finishes with scanning the obstacle, the second component of the algorithm comes into action by calculating the best position to jump from and then adjusts its stride to land without touching the barrier.

The third part of the algorithm computes the jumping trajectory of the robot as comes into action as soon as the robot finishes of with calculations for determining the best position to jump from. Based on the data from all the other components of the algorithm, this part determines the amount of force the robot’s electric motors should apply to launch the robot over the obstacle safely.

Interestingly, Kim says that the algorithms are for providing an appropriate solution, not the optimal one for the jump control.

“If you want to optimize for, say, energy efficiency, you would want the robot to barely clear the obstacle — but that’s dangerous, and finding a truly optimal solution would take a lot of computing time,” Kim says. “In running, we don’t want to spend a lot of time to find a better solution. We just want one that’s feasible.”