Programmable Chip Turns a Robot's Long Pauses into Quick Action
Duke University researchers have found a way to speed up robotic motion planning by three orders of magnitude while using one-twentieth the power. Their solution is a custom processor that can perform the most time-consuming part of the job - checking for all potential collisions across the robot's entire range of motion - with unprecedented efficiency. Motion planning for something like a robotic arm usually involves first generating a probabilistic road map, or PRM. Getting the number of edges in the PRM down to something manageable is important because of the limits of the processor that handles the planning. The new processor, a field programmable gate array, is programmed with an array of collision-detection circuits, each one of which corresponds to one of the edges in the PRM.
More From SAE Media Group
Aerospace & Defense Tech Briefs
FPGAs and HPC
Motion Design INSIDER
Controller Helps Robots Avoid Collisions
Motion Design
Microelectronics for Smarter Industrial Automation
Aerospace & Defense Tech Briefs
New Products
Aerospace & Defense Tech Briefs
SHIP Prototype Computer Chip
Automotive Engineering
Real-Time Processors Key to the Zonal E/E Revolution
Autonomous Vehicle Engineering
Daimler-Xilinx Linkup Spotlights Growing Use of FPGAs
Photonics & Imaging Technology
Optimizing Industrial Embedded Imaging Development
Robotics & Automation INSIDER
Electricity-Free Hardware Allows More Space for Robots to 'Think'
Electronics & Sensors INSIDER
LEGO-Like Artificial Intelligence Chip
Motion Control and Automation Technology INSIDER
New Technique Enables Real-Time Robotic Motion Planning
Electronics & Sensors INSIDER
Brain-Inspired Highly Scalable Neuromorphic Hardware
Embedded Technology
Digital Filtering Development Tool (Quickfilter Technologies)
Motion Control & Automation Technology
Motor Control Chips
Photonics & Imaging Technology
Rethinking FPGAs for Smart Cameras
Robotics & Automation INSIDER
Humans in the Loop Help Robots Find Their Way
Off-Highway Engineering
Shifting Design of Autonomous Architectures
Aerospace & Defense Tech Briefs
FPGA
Electronics & Sensors INSIDER
A New Design Provides the Greatest Precision in Memory to Date
Electronics & Sensors INSIDER
Advancing Semiconductor Devices for Artificial Intelligence
Transcript
00:00:00 this video shows a robot arm whose motions are planned by our special purpose processor the processor is customized hardware that we have designed specifically to perform collision detection using a PRM the processor is a collection of collision detection circuits each of which determines whether or not an edge collides with a depth pixel the circuits
00:00:20 for all edges in the PRM execute simultaneously in parallel the processor is implemented on a field programmable gate array or FPGA in this set of experiments the robot arm performs pick-and-place tasks perception is performed at runtime by 4 connect two's mounted on a wooden frame above the table which perceive the environment at 30 Hertz and transmit depth clouds to a
00:00:43 PC which combines them into an occupancy grid which is transmitted to the FPGA note that motion planning only begins after the red button is pressed as you will see motion planning is completed in less than one millisecond and there is no noticeable delay between the button press and the arm movement here we reconfigured the obstacles and the goal to highlight the motion
00:01:26 planning is being performed in real time we now show real time planning for pick-and-place with multiple goal objects again motion planning only starts when the red button is pressed