ENGS 89/90 Reports
Remote-Controlled Soft Target for Automated Vehicle Testing
Degree Program
B.E.
Year of Graduation
2019
Sponsor
Matthew Wallach Toyota Research Institute 2311 Green Rd Suite E Ann Arbor, MI 48105
Project Advisor
Laura E. Ray
Document Type
Report
Publication Date
3-1-2019
Abstract
Maneuverable, reliable, and durable soft targets are an indispensable part of autonomous vehicle test- ing. These targets allow for simulation of vulnerable road users, such as pedestrians, cyclists, or wildlife, in roadway scenarios. Our group is developing a soft target that can be a) used repeatedly in one-off testing scenarios, b) does not damage a car upon impact, c) has three interchangeable targets, and d) can be controlled remotely by human operators. The four main components of our final design are 1) a foam target, 2) an aluminum shield, 3) mechanical and electrical parts underneath the shield, and 4) a user held controller.
For our foam soft target, we used rectangular blocks of an off-the-shelf, mid-density, and closed-cell foam. These blocks are encased in polyurethane soft leather, and then further covered with polymer and rubber based adhesives (proprietary name of ‘Gorilla Tape’) to create a waterproof seal. These blocks are velcroed together so that they break apart easily upon impact with a car but can still remain upright and detectable during testing simulations. These blocks are able to be reconfigured for each of the three target configurations enumerated in our deliverables agreement – a pedestrian, a cyclist, and small wildlife. All of the targets utilize velcro as the detachment mechanism from the base as this is an inexpensive, effective, accessible, and user-friendly attachment method.
For the aluminum base, we have designed an octagonal shell with underside bulkheads that are welded together. These bulkheads serve as structural supports for the shell and provide an organizational grid for the electronic and mechanical components. Additionally, a removable underside octagonal base plate is screwed into the top shield so that the batteries and other pertinent components are separated from the environment yet still easily accessible. An additional traction pad is affixed to the bottom of the base plate to prevent skidding of our device during collisions. When run over by a car, the spring suspension system underneath the shield compresses and lowers the shield to the ground, thereby protecting the electronics underneath.
While iterating on our design throughout the term, our approach to the mechanical and electrical de- sign shifted from utilizing pre-made hobby-based systems to custom building robotics systems. While this change dramatically and significantly complicated our design, we believe this was a necessary step to de- velop an appropriate solution for TRI’s testing parameters. In our final design, we have two CIM brushed motors, a RageBridge electronic speed control, Zeee 7200 mAh 25.9 V battery, a gearbox, two powered rear castor wheels, a passive front omni-directional wheel, and an independent spring-based suspension system for each of these three wheels. To ensure safety during operation, current limiters and fuses were also wired between all of the electrical components. For our remote controller, we selected a Futaba 6J Radio System Controller since it is equipped with two programmable mixes as well as incorporating Frequency Hopping Spread Spectrum.
In total, our design costs $2,100 in materials and components. Based on financial comparisons to Toyota Research Institute’s larger scale testing soft target and industry standard models, our design is economically viable. Moving forward, we will be delivering our design specifications and prototypes to the engineers at TRI so that they can reproduce and modify our design as they desire.
Level of Access
Restricted: Campus/Dartmouth Community Only Access
Dartmouth Digital Commons Citation
Darwin, Sarah; Fett, Eric; Jennings, Alexa; Kynoch, Mackenzie; Morgan, Akaylia; Rodriguez, Karina; Uwayezu, Marie Josee; and Wang, Michelle, "Remote-Controlled Soft Target for Automated Vehicle Testing" (2019). ENGS 89/90 Reports. 3.
https://digitalcommons.dartmouth.edu/engs89_90/3
Restricted
Available to Dartmouth community via local IP address.