Why robots still can’t match animal speed and agility

Credit: Unsplash+.

Despite decades of research and millions of dollars in investment, robots still can’t match the speed, endurance, or agility of many animals.

Dr. Max Donelan from Simon Fraser University explains that animals like wildebeests, mountain goats, and even cockroaches display capabilities in natural settings that far surpass current robotic abilities.

This gap exists despite the fact that in many ways, individual robotic components are technically superior to their biological equivalents.

A team of researchers from several prestigious institutions conducted a detailed study published in Science Robotics to understand why robots lag behind.

The study, involving experts like Dr. Sam Burden from the University of Washington and Dr. Tom Libby from SRI International, compared specific subsystems used in robots—such as power, structure, and control mechanisms—with those in animals.

The results were surprising.

While it was previously believed that the biological components of animals were inherently superior, the study found that engineering parts often performed better or were on par with biological parts.

However, the real advantage for animals comes from how these parts are integrated and controlled as a whole system.

Animals have evolved over millions of years, refining their abilities to run, jump, and maneuver through complex environments with incredible efficiency. Robots, on the other hand, are relatively new inventions.

Dr. Kaushik Jayaram from the University of Colorado Boulder and Dr. Simon Sponberg from the Georgia Institute of Technology also contributed to the research, highlighting the rapid progress robotics has made in a short time compared to the slow, undirected process of natural evolution.

This comparison gives hope to roboticists. Unlike biological evolution, which cannot directly transfer learned improvements, advancements in robotics can be shared and implemented quickly across different systems.

This ability to rapidly iterate and integrate new technologies could eventually allow robots to catch up with their biological counterparts.

The potential applications for efficient and agile robots are vast. They could revolutionize how we handle ‘last mile’ delivery problems, conduct search and rescue missions in dangerous environments, and manage hazardous materials.

The challenge lies not just in improving the hardware but in better integrating and controlling what already exists.

The study’s findings emphasize the importance of learning from biological systems to improve robotic integration and control. As Dr. Donelan notes, by adopting integration principles from biology, future robots could become as efficient and capable as the animals they seek to emulate.

This research marks a crucial step toward realizing that goal, potentially leading to breakthroughs in how robots move and operate in the real world.