The DARPA Robotics Challenge Trials held in December 2013 provided a landmark demonstration of dexterous mobile robots executing a variety of tasks aided by a remote human operator using only data from the robot's sensor suite transmitted over a constrained, fieldrealistic communications link. We describe the design considerations, architecture, implementation and performance of the software that Team MIT developed to command and control an Atlas humanoid robot. Our design emphasized human interaction with an efficient motion planner, where operators expressed desired robot actions in terms of affordances fit using perception and manipulated in a custom user interface. We highlight several important lessons we learned while developing our system on a highly compressed schedule.

The DARPA Robotics Challenge Trials held in December 2013 provided a landmark demonstration of dexterous mobile robots executing a variety of tasks aided by a remote human operator using only data from the robot's sensor suite transmitted over a constrained, fieldrealistic communications link. We describe the design considerations, architecture, implementation and performance of the software that Team MIT developed to command and control an Atlas humanoid robot. Our design emphasized human interaction with an efficient motion planner, where operators expressed desired robot actions in terms of affordances fit using perception and manipulated in a custom user interface. We highlight several important lessons we learned while developing our system on a highly compressed schedule.

Successful command and control (C2) of autonomous vehicles poses challenges that are unique to the marine environment, primarily highly restrictive acoustic communications throughput. To address this, the Unified C2 architecture presented here uses a highly compressed short message encoding scheme (Dynamic Compact Control Language or DCCL) to transfer commands and receive vehicle status. DCCL is readily reconfigurable to provide the flexibility needed to change commands on short notice. Furthermore, operation of multiple types of vehicles requires a C2 architecture that is both scalable and flexible to differences amongst platform hardware and abilities. The Unified C2 architecture uses the MOOS-IvP autonomy system to act as a "backseat driver" of the vehicle. This provides a uniform interface to the control system on all the vehicles. Also, a hierarchical configuration system is used to allow single changes in configuration to propagate to all vehicles in operation. Status data from all vehicles are displayed visually using Google Earth, which also allows a rapid meshing of data from other sources (sensors, AIS, radar, satellites) from within, as well as outside of, the MOOS-IvP architecture. Results are presented throughout from the CCLNET08, SQUINT08, GLINT08, GLINT09, SWAMSI09, and DURIP09 experiments involving Robotic Marine surface craft (ASCs) and Bluefin, OceanServer, and NURC underwater vehicles (AUVs).