Self driving tech by Plus.ai
To enable large-scale commercial transport, Plus.ai has developed selfdriving trucks.
With no legacy as such in the area of commercial vehicle manufacture, US-based Plus.ai has developed self-driving trucks to enable large-scale movement of cargo. Leveraging its experience in the field of full-stack selfdriving technology, the company, founded in 2016 by a group of serial entrepreneurs and industry veterans with over 20 years of experience in high tech and artificial intelligence, is turning largescale commercialisation of autonomous transport into reality. Claiming to lead in the space of self-driving truck technology with the likes of Daimler and Tesla in the fray, the company, with its headquarters in the Silicon Valley of California, is stressing much on the use of latest automation technology. A brainchild of Stanford University graduates David Liu and Hao Zheng, Plus.ai is thus providing much impetus to research in the direction of combining the latest in automation with the best of truck tech.
Having its R&D department in China, the company is said to be gaining good traction with
leading truck manufacturers, large shippers and fleet operators on the basis of its semi-truck proto successfully hauling a refrigerator trailer full of butter across the country amounting to a trip of 2,800 miles in less than three days. Claiming to differentiate itself by working with CV makers and fleet operators rather than forward the idea of launching its own trucks like Tesla, Plus.ai has been running pilot freights in an effort to offer best-inclass level four autonomous driving systems with a safetyfirst approach. Working with CV makers in the US as well as China, the company is striving to get California’s Autonomous Testing License. It is shaping the future of trucking with the use of highly advanced autonomous technology in the interest of safety and superior fuel efficiency. Confident that businesses will turn to autonomous trucks in a big way in the future for the advantages they will offer, Plus.ai is aiming at a reduction in fuel costs by as much as 25 per cent.
Demonstrating the ability of an autonomous truck to safely navigate the Great Plains, winding roads in the Rockies, road construction, long tunnels, over 11000 feet elevation and wet and snowy roads with the 2,800 miles cross-country trip, and irrespective of whether it is day or night, Plus.ai is also highlighting the reliability of the technology that it has developed. Mentioned an plus.ai source, that their autonomous truck carried almost 40,000 pounds of Land O’Lakes butter from California to Pennsylvania, completing one of the first coast-coast autonomous truck deliveries during Christmas last year. He said that the truck not only proved that it was reliable, it also put the spotlight on efficiency gains. Carrying onboard a driver and one operation specialist systems engineer to meet the legal requirements relating to the operation of autonomous vehicles on public roads, the truck provided Plus.ai with realtime data and added much to the knowledge of further finetuning the technology.
The truck driver took to the wheel only while getting off the highway for rest stops and breaks, claimed the source. He informed that the truck was equipped with the company’s advanced autonomous driving system, which utilises multi
modal sensor fusion, visual algorithms, and simultaneous localisation and mapping (SLAM) technologies. With shortage of drivers projected as the key reason behind the development of autonomous driving technologies, Plus.ai is also looking at how such a technology could benefit in terms of road safety, efficiency and emissions. Of the opinion that autonomous technology will first find its way into the trucking space as far as its commercialisation is concerned, the source explained that there are yet many challenges to be tackled until such a scenario becomes a reality. Stating that the company will test self-driving trucks in all permissible continental states in the US until the end of 2020, the source revealed that they are in discussion with Paccar, Navistar’s International truck division and FAW to name a few. He informed that his company has also introduced a ‘Safety Testing Program’ to help companies to prepare their self-driving trucks to navigate in all terrains and weather conditions. Developing new metrics that can measure the readiness of its autonomous driving system for commercial deployment, Plus.ai is looking at launching its first commercial product by 2023.
CUTTING EDGE TECH
Developing advanced autonomous driving systems to power hub-to-hub autonomous ‘Class eight’ vehicles that are typically a part of the large fleets, Plus.ai made much use of sensors positioned at the right places. The company took into consideration the height of the heavy trucks in doing so. This enabled it to arrive at systems that would see vehicles running at a distance. Employing cameras, radar and lidar (light detection and ranging) technology, Plus.ai carried out much research in the area of computer vision software. It helped to underpin artificial intelligence. This in-turn helped
to get the truck to have a 360 degree vision. The result of this was superior operational safety, efficiency and maturity.
Capable of tracking vehicles at a distance of one mile (or 1600 m) with the help of its fusion-based perception system, the autonomous truck tech of Plus.ai includes localisation and mapping algorithms. This ensures accurate tracking and identification of the vehicle location. The map update in-turn is extremely accurate therefore. Since this is possible, it is also possible to map the truck against its surroundings rather accurately. Such technological properties could also help explain how the proto truck that ferried butter across 2,800 miles performed complex tasks like detection and analysis of ground objects and road structures extremely accurately. The tech also led to highly successful behaviour prediction of the proto truck as well as the surrounding vehicles.
Deploying a number of deep learning models and offering multiple levels of redundancy through its safetyfirst approach, Plus.ai has been successful in reducing the impact of sensor failures. Sensors are a part, but not the only ones. They are a part of the scheme to correct localised environmental noise among other components, a procedure that significantly boosts detection and analysis precision as far as the truck and its surroundings are concerned. Offering software redundancy through its complementary models and mechanisms such as odometry, VisualSLAM and PointCloudbased localisation, Plus.ai has streamlined its functions and ensured smooth transitions between different operating modes by using statistical models. By collaborating with the OEMs, the company is looking for deep integration of software systems with vehicle redundant electrical and mechanical systems. Plus. ai, in fact, carried out actuation redundancy. It successfully performed compute redundancy by completing level four autonomous system and the minimum safe landing system. This boosted the safety performance against the likelihood of a software, sensor or hardware failure.