VEMI Lab researchers earn federal prize, invite to White Home for software program that makes self-driving automobiles extra accessible – UMaine Information

The VEMI lab on the College of Maine earned third place in a nationwide competitors and an invite to the White Home for creating an inclusive smartphone software program platform that may present navigational help to individuals with visible impairment and seniors who wish to use self-driving automobiles for ride-sharing and hailing providers.  

VEMI will obtain $300,000 for profitable the prize within the second section of the U.S. Division of Transportation’s Inclusive Design Problem for its Autonomous Automobile Assistant (AVA) smartphone know-how. VEMI leads the group designing the AVA platform, generally known as the Autonomous Automobile Analysis Group (AVRG), which additionally consists of collaborators from Northeastern College and Colby Faculty. 

For its problem, the DOT sought proposals for accessible and inclusive design options that may assist individuals with disabilities use autonomous autos for employment and important providers. VEMI was invited to take part in Stage II of the problem after being named a semifinalist within the first section.  

First- and second-place winners within the competitors had been Purdue College and AbleLink Good Residing Applied sciences, respectively.

The prizes within the Inclusive Design Problem had been introduced July 26 as a part of the DOT’s celebration for the thirty second Anniversary of the People with Disabilities Act, generally known as the ADA. VEMI director Richard Corey and chief analysis scientist Nicholas Giudice are taking part in a ceremony on the division’s headquarters in Washington, D.C., adopted by a coverage session hosted by the Workplace of Public Engagement and Workplace of Science and Expertise Coverage July 27 on the White Home. 

“That is thrilling nationwide recognition of one of many excellent, ongoing analysis and growth initiatives from our VEMI Lab,” says UMaine President Joan Ferrini-Mundy. “Such innovation addressing wants, together with these for fairness and inclusion, and offering important developments in know-how are amongst our R1 analysis college achievements.”

“The entire VEMI Lab crew is ecstatic to obtain the help and nationwide recognition for our innovation,” Corey says. “This prize will assist help our college students who’re main the cost in human automobile collaboration analysis and accessible know-how at VEMI. Working with the IDC crew on the U.S. DOT has been a delight and we’re deeply honored to have been chosen as high tier winners.” 

Designing the undertaking and incomes the prize for it was a collaborative effort from VEMI workers, college students and exterior companions, particularly Grant Beals, Paul Fink, Aubree Nygaard, and Raymond Perry from VEMI; Xue (Shelley) Lin, Pu Zhao and Yushu Wu from Northeastern College; and Stacy Doore and Matthew Maring from Colby Faculty. 

AVA will assist customers request a automobile, discover it, enter it, exit it and journey to their chosen vacation spot. It gives a multisensory interface that gives customers steering via audio and haptic, or touch-based, suggestions and high-contrast visible cues. To supply the performance, researchers utilized GPS, LiDAR, Gyroscope and Accelerometer know-how; real-time pc imaginative and prescient by way of the smartphone digital camera; machine studying; synthetic intelligence and different software program. 

Customers will create a profile in AVA that displays their accessibility wants and present strategies of navigation so the software program can discover appropriate transportation for them. When the automobile arrives, AVA will information the person to it utilizing the digital camera and augmented actuality (AR), which gives an overlay of the setting by superimposing high-contrast strains over the picture on the smartphone display to focus on the trail, and verbal steering resembling compass instructions, road names, addresses, close by landmarks and different indicators. The software program additionally will pinpoint environmental hazards, together with low-contrast curbs, site visitors cones and overhanging obstructions like branches and man wires, by emphasizing them with contrasting strains and vibrations when customers method them. It would then assist customers discover the door deal with to enter the automobile awaiting them. It additionally makes use of the identical capabilities to assist the person when exiting their automobile to seek out their vacation spot. 

AVA will supply accessible modules with simulations that practice customers not solely tips on how to use the applying, but in addition coaching on tips on how to work together with journey sharing and hailing providers with self-driving autos when an individual is now not out there to supply help. For future initiatives, researchers plan to develop extra software program that may permit riders to make use of the know-how to attach with the automobile management programs whereas using. These instruments will embrace multisensory maps, context-aware gesture interactions and software programming interfaces, all of which can help in-cabin accessibility. 

Watch the video presentation about AVA ready by VEMI Lab researchers for the DOT’s Inclusive Design Problem to study extra. 

“Autonomous autos have the potential to be a really recreation altering, disruptive know-how for enhancing accessible, inclusive transportation for individuals with visible impairments and older adults,” says Giudice, additionally a UMaine professor of spatial computing and congenitally blind. “Nonetheless, to succeed, there are a variety of challenges to beat first. Our preliminary analysis and growth of AVA within the first IDC semi-finalist spherical has made vital progress in addressing present limitations, however I’m most enthusiastic about our future growth made doable by this finalist IDC prize, which can result in a strong, end-to-end inclusive journey answer that integrates with different accessible apps and platforms.” 

The AVA undertaking builds on a Nationwide Science Basis grant led by Giudice and Corey on belief constructing and human-vehicle collaboration with autonomous autos, in addition to a seed grant-funded, joint effort between UMaine and Northeastern College to enhance accessibility, security and situational consciousness inside self-driving autos. Analysis on each initiatives goals to develop a brand new mannequin of human-AI automobile interplay to make sure individuals with visible impairments and seniors can higher perceive what their autonomous automobile does throughout their travels, and so the autos can successfully talk with them — work that will probably be instrumental for informing future AVA growth on this Inclusive Design Problem prize. 

AVA serves as one instance of the broad AI, computing and knowledge programs analysis VEMI scientists and others are conducting at UMaine. Their work exemplifies the analysis and public service missions of the top-tier R1 college, a designation UMaine earned earlier this yr from the Carnegie Classification of Establishments of Larger Schooling. 

VEMI, co-founded by Corey and Giudice in 2008, explores completely different options for fixing unmet challenges with know-how. Prime areas of analysis and growth pertain to self-driving autos, the design of bio-inspired instruments to enhance human-machine interplay and performance, and new information-access know-how to enhance inclusive environmental consciousness, spatial studying and wayfinding for each sighted and visually impaired navigators.  

Contact: Marcus Wolf, 207.581.3721; [email protected]