How autonomous robots are learning to navigate environments once considered impossible for machines
From the dusty plains of Mars to the dark, submerged tunnels of hydropower dams, robots are evolving. By integrating risk-aware decision making, advanced sensory skins, and dual-medium mobility, engineers are finally bridging the gap between rigid automation and the fluid, intuitive movement required to explore our most inaccessible frontiers.
The quest for true autonomy is pushing robotics into extreme environments. On Mars, the Perseverance rover, which landed in February 2021, utilizes the most sophisticated autonomous driving capabilities ever deployed on the planet. This technology is vital for its mission to uncover signs of ancient life, relying on machine learning and risk-aware decision making to navigate the Martian surface safely and efficiently without constant human intervention.
On Earth, researchers at the Shanghai Research Institute for Intelligent Autonomous Systems (SRIIAS) are tackling the limitations of traditional sensory awareness. Led by Bin He, the team has developed innovations like the 'TJ-FlyingFish,' a drone capable of transitioning between air and water. By utilizing a dual-gear system and swivel-mounted rotors, it overcomes the dense drag of water that typically disables aerial drones. This allows it to monitor water quality and inspect infrastructure, such as long dam tunnels, that were previously unreachable by conventional equipment.
Simplicity and energy efficiency remain the primary hurdles for scaling these technologies. To address the power-intensive nature of wireless sensor networks, the SRIIAS team designed algorithms that compress data on-board, allowing robots to make real-time decisions in the field. They have also created a low-power, Venus flytrap-inspired grasper that snaps shut in 0.18 seconds, enabling robots to perch on natural surfaces like tree branches or cave walls for long-term monitoring. By prioritizing modular, low-energy designs, these engineers aim to move robotics out of the laboratory and into practical, scalable applications for disaster response, ecological conservation, and remote infrastructure maintenance.
Source: To Boldly Go Where No Robots Have Gone Before: Exploration with Autonomous Robots (Live Public Talk)
