Pothole Detectives
Vehicles traveling repeated routes are data-gatherers for Aisin’s Road Maintenance Solution (ARMS) technology, ready for beta testing in the U.S.
The first North American business-to-government technology initiative for Aisin’s Connected & Sharing Solutions Division is expected to begin beta testing in late 2023 or early 2024. “ARMS (Aisin Road Maintenance Solution) technology is a system that can identify potholes and other road problems, then communicate those issues from the vehicle to the cloud in real time,” Scott Turpin, president and CEO of Aisin World Corp of America, said in an SAE Media interview during the 2023 North America International Auto Show in Detroit.
Vehicles that are routinely driven on the same roads within a community are well-suited for ARMS application. “Garbage trucks, buses, snowplows and other vehicles that run similar routes on a regular basis present a way to capture road data that can then be accessed in the cloud by a municipality,” said Cam Danh, Assistant Manager, Business Design Group of Aisin’s Connected & Sharing Solutions Division.
A vehicle equipped with a high-definition camera and an edge computer is required hardware for ARMS. Aisin’s location-based services platform, which relies on machine learning and algorithms, enables the pinpointing of a road issue. Initial data processing occurs via the onboard edge computer with additional analysis in the cloud. Edge computing is an information architecture that brings the computing resource closer to the source of data.
The ARMS system can detect the depth of potholes, various types of road cracks and road roughness. “Digitalization of the collected data means a municipality will have a library of road information, including historical reference points,” Danh said. The collection of road data will occur year-round, as the ARMS-equipped vehicles in a municipality program are driven daily or weekly.
Aisin developed and first-deployed ARMS in Japan, so the technology’s artificial intelligence (AI) model had to be trained for North American applications. “Japanese roads have different characteristics from roads in North America. If we applied the same AI model that was developed in Japan to the U.S., there would be a lot of false positives, or incorrect characterizations,” said Danh. Road differences between Japan and the U.S. include lane widths and unique manhole-cover designs.
Training the AI model for North America use has taken more than two years, according to Kevin Smith, Senior Manager, Business Development Department of Aisin’s Connected & Sharing Solutions Division. “It’s important to make sure that we have the most accurate AI model possible, so we’ve accumulated thousands of miles driving around Texas and Michigan training the model, which gets more accurate with more training,” Smith said.
Top Stories
INSIDERElectronics & Computers
Army Launches CMOSS Prototyping Competition for Computer Chassis and Cards
INSIDERSoftware
The Future of Aerospace: Embracing Digital Transformation and Emerging...
ArticlesAerospace
Making a Material Difference in Aerospace & Defense Electronics
INSIDERRF & Microwave Electronics
Germany's New Military Surveillance Jet Completes First Flight
ArticlesAerospace
Microchip’s New Microprocessor to Enable Generational Leap in Spaceflight...
EditorialConnectivity
Webcasts
Power
Phase Change Materials in Electric Vehicles: Trends and a Roadmap...
Automotive
Navigating Security in Automotive SoCs: How to Build Resilient...
Automotive
Is Hydrogen Propulsion Production-Ready?
Unmanned Systems
Countering the Evolving Challenge of Integrating UAS Into Civilian Airspace
Power
Designing an HVAC Modeling Workflow for Cabin Energy Management and XiL Testing
Defense
Best Practices for Developing Safe and Secure Modular Software