Carnegie Mellon University’s Robotics Institute is spearheading groundbreaking research to develop the next generation of explorers – robots capable of autonomous exploration. The Autonomous Exploration Research Team has successfully created a suite of robotic systems and planners that enable robots to explore unknown environments more efficiently, probe uncharted areas, and generate accurate and detailed maps, all without human intervention.
The team’s innovative systems empower robots to autonomously navigate and map their surroundings, making the exploration process seamless and efficient. Ji Zhang, a systems scientist in the Robotics Institute, explained that these robots can be deployed in various environments, such as department stores or disaster-hit residential buildings, and sent on their exploration mission without human guidance. In real-time, these robots construct detailed maps while determining their next path of exploration, allowing users to observe the entire environment without stepping into the space themselves.
Over the past three years, the research group has tested their exploration systems in diverse locations, including underground mines, parking garages, the Cohon University Center, and various indoor and outdoor areas on the CMU campus. The team’s systems can be integrated into a wide range of robotic platforms, effectively transforming them into modern-day explorers. For testing purposes, the group employs a modified motorized wheelchair and drones.
The robots equipped with the exploration systems operate in three different modes. In the first mode, a human operator can control the robot’s movements while the autonomous systems prevent collisions. In the second mode, the operator can select a point on the map, and the robot will navigate to that specific location. The third mode is complete exploration autonomy, where the robot independently explores the entire space, generating a comprehensive map.
Howie Choset, a professor in the Robotics Institute, emphasized the system’s adaptability for various applications, from delivery services to search-and-rescue missions. The team combined a 3D scanning lidar sensor, a forward-looking camera, and inertial measurement unit sensors with an exploration algorithm, enabling the robot to determine its location, track its movements, and plan its next route. The resulting systems are significantly more efficient than previous approaches, producing more complete maps while reducing the algorithm run time by half.
The versatility of these new systems allows them to perform exceptionally well in challenging conditions with low-light and spotty communication, such as caves, tunnels, and abandoned structures. The exploration system developed by the group powered Team Explorer, a collaborative entry from CMU and Oregon State University in DARPA’s Subterranean Challenge. Team Explorer secured fourth place in the final competition and won the Most Sectors Explored Award for mapping more of the route than any other team.
The team is committed to open-sourcing all their work, aiming to empower society with the capabilities of building autonomous exploration robots. Chao Cao, a Ph.D. student in robotics and the lead operator for Team Explorer, highlighted the significance of this fundamental capability, stating that once it is achieved, a wide array of possibilities opens up for robotics and exploration endeavors.