The humble vehicle, purpose-built to carry people and goods from one place to another, is rapidly transforming into a “computer on wheels,” as Weisong Shi, professor of computer science at Wayne State University describes it. With advanced cameras and sensors, connected and autonomous vehicles will communicate with each other and with the infrastructure to make transportation safer and more efficient.
In sensing and analyzing the environment, however, future vehicles are predicted to generate huge amounts of data. Some researchers predict that one vehicle could create between four and 10 terabytes of data each day.
To study the management of vehicle data and develop the most effective processing and analysis methods, Shi has established the Connected and Autonomous Driving Laboratory (CAR Lab
) at Wayne State University. The lab will work with industry and public sector partners to “design and implement enabling technologies” to meet the data challenges and realize the opportunities created by connected and autonomous driving.
MetroCAD is an annual gathering for industry research presentation.
Intelligent vehicles require real-time processing of pedestrian-detection data, a task which challenges the model of cloud-computing, so Shi and his colleagues are working on edge computing solutions, which, as its name implies, keeps data at the edge of the network. “This computing has to be done on the vehicle, or maybe in the infrastructure,” says Shi, who is director of the CAR Lab.
Under normal 4G LTE conditions, it would take 18 days to transmit the amount of data generated by a connected vehicle in just one day. “Connected and autonomous vehicle technology is the perfect application for edge computing,” Shi says. The CAR Lab brings together university research talent on edge computing, vehicle-to-vehicle (V2V) communications, and cyber-physical systems, and collaborates with industry innovators to develop and test advanced technologies.
Industry advisors support innovation
One such innovator is California- and China-based PerceptIn
. With a goal of bringing affordable low-speed, last mile mobility solutions to market, PerceptIn develops visual intelligence products. Initially attracted to the CAR Lab for its high-level expertise in the center of the Detroit automotive industry, PerceptIn plans to commercialize its technology in the Detroit region, which is the geographic location that makes the most sense.
“Silicon Valley is very good at cutting-edge R&D, but the automotive industry global center is Detroit. You can do R&D in California or in China, but if you are commercializing, then Detroit is the place to be,” says Shaoshan Liu, PerceptIn president and co-founder. “We want to enable the CAR Lab to do more cutting-edge research in the automotive city.”
Another CAR Lab industry partner is working to bring mobility solutions to market sooner rather than later. Through a unique, patented, low cost per mile system, Ann Arbor-based Bedestrian
will roll out autonomous delivery in robotic-like vehicles that employ an advanced, natural language interface. This combination will be game-changing for those with limited mobility, like the elderly, says Shadi Mere, CEO of MaverickAI, the innovator behind Bedestrian.
“We can help retail pharmacies deliver to urban settings the same day, and the voice will speak naturally to the user. Theoretically, an individual can visit their doctor, and have the prescription at the doorstep when they get home,” Mere says.
In developing this technology, MaverickAI looked at solutions they could provide next year, not in five years, when autonomous vehicles are expected to roll out. And Detroit, with its already-established mapping and navigation innovation, was the right place to do it. “Health and mobility have big futures in the Michigan economy, and my team is Michigan-based,” he says. “There is a huge wealth of talent here in Michigan...people forget that patents and IP all start here.”
Solutions like Bedestrian value the deep knowledge that sprout from academic environments, and are excited to work with the CAR Lab to further innovate. “Academia is very helpful. We do things like machine vision, and other challenges that need the latest white papers that are being published,” says Mere. “In this mobility age, academia will play a big role.”
Open platform solutions
One goal for the CAR Lab is to develop an open vehicular data analytics platform (OpenVDAP) which will allow for third-party innovation to better meet the needs of end users and stimulate tech innovation.
“Maybe someone can develop an algorithm to give better early warnings than the original OEMs [have developed],” says Shi. “When you open the platform to the public, some will see a great opportunity for startups to create solutions.”
Ultimately, the possibilities for application development are endless, which makes industry collaboration so valuable.
Earlier this year, the CAR Lab hosted MetroCAD
, an annual workshop event to allow industry developers to present their research. More than 70 individuals from local industry and academic institutions, including MIT, attended.
“We will continue this workshop next year,” says Shi. “It’s the perfect place to serve as a synergy between innovation in [Wayne State] and other universities. People can come here to present their research and results and OEMs and Tier1 suppliers and startups can attend to look for potential partnerships.”