A new platform being developed by connect vehicle firm Wejo is to drive autonomous vehicles forward.
The North West-headquartered company, which is listed on the NASDAQ in New York, will reveal its Neural Edge processing platform with partner Microsoft at CES in Las Vegas.
It says the trademarked platform uses machine learning to address data overload and deliver faster, more cost-effective and sustainable vehicle communication insights.
Wejo, which processes 16 billion data points from 11 million vehicles a day and has a heavy presence in the US, placed third on our TransportTech 50 innovation ranking late last year.
With so much rich data coming from vehicles today, latency and data storage costs are potential obstacles in harnessing and scaling the power of real-time vehicle communications – both with other vehicles and the infrastructure that is set to power smart cities.
Neural Edge – which leverages Wejo’s strategic partnership with Microsoft Azure and is powered by Wejo’s ADEPT platform – optimises how this data is managed within the vehicle, further processes it at the edge and ultimately communicates to the cloud.
The process aims to support safer vehicles, enable further advancements in electric vehicle and autonomous mobility, and reduce congestion and emissions.
Wejo recently revealed rocketing customer numbers and revenues.
“When I started Wejo in 2014, I knew that the proliferation of new mobility technology would drive data to a tipping point. And we are at that point today,” said Richard Barlow, founder & CEO.
“With today’s vehicles producing approximately 25 gigabytes of data per hour, and as vehicle technology advances adding more sensors, data filtering and neural edge processing technology is essential to reduce this overload and drive the industry forward.
“Partnering with Microsoft and Palantir has positioned us to address this problem today, and to look ahead at the benefits of Wejo Neural Edge as a driver in the growth of autonomous mobility.”
Wejo Neural Edge will filter and analyse vast amounts of AV, EV and CV data before transmitting only the essential information to the cloud.
By utilising machine learning algorithms to reconstruct vehicle journey and event data, it can also take 20% of the data and reconstruct it to represent 100% of the data, without any loss in fidelity or integrity. This has positive implications for the environment due to reduced data storage needs.
It will also deliver digital twins of vehicles and cities to reshape how we view the entire product and service ecosystem related to mobility. In a simulation environment, a digital twin of the US can be constructed to simulate how vehicles in different cities need to respond and navigate without having to outlay massive infrastructure costs of physical hardware or vehicles to be able to relearn how a vehicle should behave.
“At Wejo, we believe that digital twins will reshape everything from road safety, to insurance, advertising, after-sales and more,” said David Burns, CTO.
“With Wejo Neural Edge we can look at what a CV is doing a kilometer away, and then alter and change the driver experience of an AV based on the information that is coming from down the road.”