Engineering news
The new planning algorithm for robots with arm-like appendages was developed at the University of Michigan. It found successful paths three-times as often as standard algorithms, the developers said, with much less processing time.
“In a collapsed building or on very rough terrain, a robot won't always be able to balance itself and move forward with just its feet,” said associate professor Dmitry Berenson. “You need new algorithms to figure out where to put both feet and hands. You need to coordinate all these limbs together to maintain stability, and what that boils down to is a very difficult problem.”
The research enables robots to determine how difficult terrain is before calculating a successful path forward, which might include bracing on a wall with one or two ‘hands’ while taking the next step forward.
“First, we used machine learning to train the robot on the different ways it can place its hands and feet to maintain balance and make progress,” said PhD graduate Yu-Chi Lin. “Then, when placed in a new, complex environment, the robot can use what it learned to determine how traversable a path is, allowing it to find a path to the goal much faster.”
Even using that estimate, traditional planning algorithms can take a long time. "If we tried to find all the ‘hand’ and ‘foot’ locations over a long path, it would take a very long time,” said Berenson.
The team used a ‘divide and conquer’ approach instead, splitting a path into tough-to-traverse sections, where they could apply their learning-based method, and easier-to-traverse sections, where a simpler path planning method works better.
“That sounds simple, but it's really hard to know how to split up that problem correctly, and which planning method to use for each segment,” said Lin.
To do this, robots need a geometric model of the entire environment. This could be achieved in practice with a flying drone that scouts ahead, the researchers said.
In a virtual experiment with a humanoid robot in a corridor of rubble, the team's method outperformed previous methods in both success and total planning time – important when quick action is needed in disaster scenarios.
Over 50 trials, the approach reached the goal 84% of the time, compared to 26% for the basic path planner. It took just over two minutes to plan, compared to over three minutes for the basic path planner.
The team also showcased their method's ability to work in the real world, using a wheeled robot with a ‘torso’ and two arms. With the base of the robot placed on a steep ramp, it used its hands to brace itself on an uneven surface as it moved. The robot used the team's algorithm to plan a path in just over a tenth of a second, compared to over three-and-a-half seconds with the basic path planner.
In future work the team hopes to incorporate dynamically stable motion – similar to the natural movement of humans and animals – which would free the robot from having to be constantly in balance, and could improve its speed.
The research was published in Autonomous Robots.
Want the best engineering stories delivered straight to your inbox? The Professional Engineering newsletter gives you vital updates on the most cutting-edge engineering and exciting new job opportunities. To sign up, click here.
Content published by Professional Engineering does not necessarily represent the views of the Institution of Mechanical Engineers.