A groundbreaking research group within Carnegie Mellon University’s Robotics Institute is revolutionizing exploration with their cutting-edge robotic systems. The Autonomous Exploration Research Team has developed a suite of autonomous robots and planners that enable advanced exploration, rapid mapping, and accurate navigation through unknown environments—all without the need for human intervention.

Led by Ji Zhang, a systems scientist at the Robotics Institute, the team’s robotic systems boast incredible versatility. The robots can be deployed in various environments, ranging from department stores to disaster-stricken residential buildings, to explore and map in real-time. This autonomous approach eliminates the need for human presence on-site, making exploration more efficient and safe.

Over the course of three years, the team has meticulously crafted exploration systems that have successfully mapped underground mines, a parking garage, the Cohon University Center, and various indoor and outdoor locations across the CMU campus. By equipping nearly any robotic platform with their cutting-edge computers and sensors, the team transforms these machines into modern-day explorers. In their testing, the researchers have utilized a modified motorized wheelchair and drones, among other platforms.

The robots operate in three distinct modes, making the system highly adaptable to various applications. In the first mode, a person can control the robot’s movements while autonomous systems ensure collision avoidance. In the second mode, users can select points on a map, and the robot navigates to those specified locations. The third mode, which is pure exploration, allows the robot to autonomously investigate the entire space and create a detailed map.

The team’s exploration algorithm, combined with a 3D scanning lidar sensor, forward-looking camera, and inertial measurement unit sensors, equips the robot with comprehensive knowledge of its surroundings. This information empowers the robot to understand its current position, track its previous movements, and determine its next course of action. Compared to previous approaches, the team’s systems are remarkably efficient, producing more comprehensive maps while significantly reducing algorithm run time.

An impressive aspect of the system is its capability to operate in low-light and challenging conditions with limited communication, such as caves, tunnels, and abandoned structures. Team Explorer, an entry from CMU and Oregon State University in DARPA’s Subterranean Challenge, was powered by a version of the group’s exploration system. The team placed fourth in the final competition and received the prestigious Most Sectors Explored Award for mapping a larger route than any other team.

Passionate about sharing their advancements, the team has made all their work open-source, aiming to strengthen society with the possibilities of autonomous exploration robots. Chao Cao, a Ph.D. student in robotics and the lead operator for Team Explorer, emphasizes that this fundamental capability opens the door to a multitude of possibilities, unlocking the potential for a wide range of applications.

By Impact Lab