Mobile Autonomous Vehicle Obstacle Detection and Avoidance

Mobile equipment is heavily used in manufacturing. There is a growing acceptance of either partially or fully autonomous equipment in the manufacturing area. A major problem, however, is that (especially small) manufacturing facilities frequently operate with people and mobile equipment moving through the same cluttered and constantly-changing environment. Safety is of paramount concern, and standards are essential to reduce the potential for injury.  The ability to control multiple autonomous vehicles from different manufacturers and to mount different sensors from different vendors is also a challenge. This project will develop standard test methods and performance measures using advanced sensor and control systems and operator alerts to help improve standards for semi-autonomous and autonomous industrial vehicles.

Develop and deploy the measurement science and standards to enable humans and autonomous mobile vehicles to work safely in close proximity on the factory floor by 2014.

Automated guided vehicles (AGV’s) are typically used in confined areas away from human interaction.  Recently, the ANSI B56.5 standard was improved with substantial input from NIST to allow non-contact sensing of specific test pieces for safe AGV’s.  Still missing are dynamic 3D measurements to improve the test methods in the standard and allow safe AGV use near humans, on flexible assembly lines, and in unstructured environments. Also, OSHA accident statistics show that high accident rates for manned forklifts are mainly due to the existence of blind spots around the vehicles.  There are over one million manned vehicles in factories, causing, on average, an accident every 3 days of which 80% include a pedestrian.  Today’s advanced 3D imaging and RFID technology and control has the potential to eliminate blind spots and to provide semi-autonomous vehicle control for safe slow-downs and stops.  The ANSI B56.5, B56.1 and B56.11.6 committees have all expressed interest in measurements that improve these standards to address visibility hazards.

The new technical idea is to develop the measurement science that will allow test methods for AGV bumper force on humans and advances in 3D imaging and non-line-of-sight (e.g., RFID) sensor technology to improve control capabilities for safe autonomous industrial vehicles.  This idea will also serve the much larger industry of manned manufacturing vehicles that will have semi-autonomous functionality. The project will develop performance measures and new test methods for sensor-based vehicle control.  It will support standards including ANSI B56.5 (part A - AGV’s and part B - automated functions of manned vehicles) and B56.11.6 (visibility for vehicles) and the recent international ISO/FDIS 13564-1 (visibility for vehicles) standard.

The B56.5 committee is requesting new work, such as further safety test methods and AGV bumper force measurement and justification.  In FY13, the project will develop performance measures to ensure safety of autonomous navigation in close proximity to moving standard test pieces and mannequins.  For example, in FY12, the NIST (surrogate) AGV^[1] was used to measure performance of mandatory emergency stop braking as compared to controlled braking on long, straight paths with preliminary results showing the need for controlled braking.  In FY13, curved paths and instantaneous changes to AGV direction will be tested through use of an industrial AGV.  Paths that allow AGVs to detect obstacles, understand lane changing, and drive-around detected obstacles will be designed and tested.  The B56.5 committee suggested that measurements be performed and test methods designed to determine the amount of force an AGV bumper can apply to a human. In FY13, a design will be developed for a standard apparatus that performs bumper force testing using standard B56.5 test pieces and fixtures. This work will be coordinated with the Safety of Human-Robot Systems in Fixed Workcell Environments project, and some of the equipment developed by that project will be used.

In FY12, a CRADA was established with a safety RFID system manufacturer, the RFID system was integrated with forklift deceleration control and preliminary tests were performed.  In FY13, this effort will be completed and findings reported to the B56.5 committee for use in predicting pedestrian intent to enter the path of a forklift or AGV from a region that is not in the line of sight of the vehicle.  In FY12, we provided a grant to Southwest Research Institute to implement a human tracking algorithm using various 3D imaging sensors for use in manufacturing environments.  In FY13, the system will be integrated onto the NIST AGV and/or forklift as a plug-and-play module to perform pedestrian detection and tracking.  Sensor data will be evaluated for autonomous and semi-autonomous control of vehicle speed to develop standard test methods for safely slowing and stopping a vehicle when near humans.  Findings will be published and provided as recommendations to standards committees.

In future years, ANSI B56.x and ISO 13564-1 standards efforts are expected to continue from the FY12 and FY13 basis.  Specifically, in support of B56.5 and B56.11.6, manned vehicle blind-spot measurement and control methods will be improved and published in a journal article and proposed to the ANSI B56.5 standards committee.

• Outcome: An improved B56.5 draft standard with non-contact sensing clarification and addition of a new test piece and test piece reflectivity properties was published in FY12.
• Outcome: NIST proposal to the B56.11.6 committee to incorporate other test methods in the current draft standard, including the use of advanced 3D measurement systems.
• Output: A paper on AGV safety was presented at a conference:
• • Bostelman, Shackleford, Cheok, Norcross: ‘Standard Test Procedures and Metrics Development for Automated Guided Vehicle Safety Standards’ PerMIS 2012
• Output: A paper was presented at the International Material Handling Research Colloquium and included as a Book Chapter:
• • Bostelman, Shackleford, Cheok, Saidi, ‘Safe Control of Manufacturing Vehicles Research Towards Standard Test Methods,’ International Material Handling Research Colloquium (IMHRC 2012).
• Output: Report from contractor (Direct Dimensions) on advanced measurement methods for forklift operator visibility.

• Project members contribute data and draft test methodologies to ANSI/ITSDF B56.5 committee for AGV’s and unmanned industrial vehicles with automated functions to ensure safe manufacturing vehicles.
• Project members contribute data and draft test methodologies to ANSI/ITSDF B56.1 and B56.11.6 committees for visibility and lift vehicles to advance the standard through improved measurement methods.

[1] NIST uses a mobile research platform as a surrogate for an industrial AGV. An AGV is being procured.