DARPA Robotics Challenge Team and Approach
Lockheed Martin Advanced Technology Laboratories successfully leads Team TROOPER with teammates from the University of Pennsylvania and Rensselaer Polytechnic Institute. Our University partners support technologies related to grasp planning, walking path planning, perception, and whole body control. Rensselaer Polytechnic Institute team members focus on upper body planning and manipulation. University of Pennsylvania team members work on perception and lower and whole body control for walking and balancing tasks. Lockheed Martin ATL’s robotics experts help bring together hardware and software, working across disciplines from mechanical and electrical engineering to computer and materials science.
In April 2015, Team TROOPER invited Lockheed Martin employees to vote to choose a new name for our robot. Employees chose “Leo” in tribute to Leonardo da Vinci, who is credited with designing the humanoid robotic knight and whose work, much like robotics, involved the blending of many disciplines.
For the DARPA Robotics Challenge, Team TROOPER is creating a software solution that controls a robot using Human Guided Autonomy, which allows a robot to primarily work on its own, contacting a human only when it needs assistance. We’re using the DARPA-furnished Atlas robot as a hardware test bed for this Human Guided Autonomy software—in other words, we’re building the “brains” of the robot using advanced perception, processing and planning technologies.
As with any new technology, these robotic systems must perform reliably, predictably, and repeatedly—especially in unexpected circumstances such as disaster prevention and relief. Our goal is to achieve collaboration between a robot and its human operators, letting humans do what humans do best with the help of robots when situations become dangerous. We envision a future where robots could even help stop a situation before it becomes a disaster by completing small intervention tasks. Opening a valve to release hot air, for example, could alter the course of events in a big way.
How We Got To The Finals
Phase 1: Virtual Challenge
The DARPA Robotics Competition is a three phase challenge. The first phase, a Virtual Challenge in June 2013, had teams from around the world complete a virtual obstacle course with a virtual robot. Team TROOPER proved successful in getting its virtual robot to drive a vehicle, walk across difficult terrain, and attach a hose connector to a spigot and turn a valve. We were one of seven teams selected to receive an Atlas robot for use in the next phases of the DRC.
Phase 2: Challenge Trials
Phase 2 was the two-day DARPA Robotics Challenge Trials on December 20 and 21, 2013, in Homestead, Florida. For this challenge, 16 teams guided their robots through eight individual physical tasks that tested mobility, manipulation, dexterity, perception, and operator control mechanisms. The TROOPER team completed tasks with their Atlas robot that included removing debris blocking an entryway, climbing an industrial ladder, and opening a door to enter a building, finishing in the top 8 to secure a spot in the Finals.
Phase 3: Challenge Finals
For the DRC’s third and final phase, Team TROOPER will be using an upgraded Atlas robot that’s 75 percent new. This updated Atlas robot is 6’2” tall and weighs 400 pounds, which is two inches taller and 70 pounds heavier than the robot used in the Trials. While the tasks performed in the DRC Finals will be similar to those of the Trials, all tasks will be performed at one time in a single series. Pushing the technology further, robots will not be connected to power chords or fall arrestors.
Did You Know?
Lockheed Martin Advanced Technology Laboratories worked with a consortium led by the University of Pennsylvania with Lehigh University to build an autonomous vehicle to compete in the DARPA Urban Challenge Competition in November 2007.
The group formed the Ben Franklin Racing team, and out of 89 teams that started the competition, they were one of only 11 teams selected to compete in the final competition.
The team was one of only six to successfully finish the course