We present the design, integration, and evaluation of a full-stack robotic system called RoMan, which can conduct autonomous field operations involving physical interaction with its environment. RoMan offers autonomous behaviors that can be triggered from succinct, high-level human input such as “open this box and retrieve the bag inside.” The robot’s behaviors are driven by a set of planners and controllers grounded in perceptual reconstructions of the environment. These behaviors are articulated by a behavior tree that translates high-level operator input into programs of increasing sensorimotor expressiveness, ultimately driving the lowest-level controllers. The software system is implemented in ROS as a set of independent processes connected by synchronous and asynchronous communication, and distributed across two on-board planning/control computers. The behavior stack drives a novel platform consisting of a pair of custom, 500 Nm/axis manipulators mounted on a rotatable torso aboard a tracked platform. The robot’s head is equipped with forward-looking depth cameras, and the arms carry wrist-mounted force-torque sensors and a mix of three- and four-finger grippers. We discuss design and implementation trade-offs affecting the entire hardware-software stack and high-level manipulation behaviors. We also demonstrate the applicability of the system for solving two manipulation tasks: 1) removing heavy debris from a roadway, where 64% of end-to-end autonomous runs required at most one human intervention; and 2) retrieving an item from a closed container, with a fully autonomous success rate of 56%. Finally, we indicate lessons learned and suggest outstanding research problems.

Over the past decade, robotics technologies and the tools used to develop them have undergone significant advancement and transformation. In this paper, we observe and assess these changes from the perspective of a 10-year research program sponsored by the DEVCOM Army Research Laboratory, named the Robotics Collaborative Technology Alliance. Beyond advancing the state of the art by conducting research at some of the top academic institutions across the United States, the alliance also worked with top government and industry partners to integrate the research into meaningful experiments and demonstrations with military relevance. This paper assesses and provides insight into the effectiveness of the collaboration tools used by the team, management methods, data collection efforts, and live and virtual experiments. Ultimately, we seek to inform future efforts requiring disparate and distant teams of the potential advantages and challenges of using such tools by providing our lessons learned for how most effectively to work as a team of teams for advancing robotics.

This paper presents the design, control, and initial performance from two iterations of human-scale (∼75 kg) quadrupedal robots built under the U.S. Army Research Laboratory (ARL) Robotics Collaborative Technology Alliance (RCTA) LLAMA (Legged Locomotion and Movement Adaptation) project. These all-electric, quadruped robots are designed with custom quasi-directdrive actuators powering 3-DOF, serial-parallel legs. To our knowledge, this is the first all-electric quadruped robot of this mass scale. The centralized energy management system uses a capacitor bank to supply burst loads and buffer regenerated energy. A hierarchical control scheme enables rapid motions (up to 1.8 m/s) over a variety of terrains. The onboard sensing suite enables deliberate, autonomous operation across rubble fields. In addition, we report on practical observations, lessons learned from field testing of two generations of the platform, and current drawbacks, such as low absolute payload (9 kg) and battery life (35 minutes). These lessons include strategies to address secondary effects at larger scales and parameters with the most impact to improve future designs.