Autonomous Vehicles Tech Digest - December 2016

Starship’s delivery robots to work with Mercedes vans

Starship Technologies, builder of delivery robots, has announced that it is partnering with Mercedes Benz vans to develop ‘Robovan’. This semi-autonomous transportation system will have Mercedes-Benz Sprinter vans acting as a mother ship for up to eight of Starship’s robots. The vans drive to delivery areas, stop at a designated place, release the robots to deliver their goods and wait for them to autonomously return. The delivery robot delivers the goods directly from a store or hub at a time the client, via a mobile app, requests. The journey from store to door can be monitored by the consignee who is alerted to the robot’s arrival by a mobile phone push notification. Starship says that it has solved the last-mile problem (the last stage of transporting goods from a transportation hub to their final destination). Mercedes Benz says this system will increase the efficiency of delivery by an order of magnitude. 
Starship’s robots have covered more than 10,000 miles in their first year since the company was founded. Starship aims to deliver goods locally in 15-30 minutes within a 2-3 mile radius. Starship has worked on commercial deliveries with Hermes, JustEat, Metro Group, Swiss Post, and Mercedes Benz. 

Changes to Google’s self-driving car project

Google’s autonomous car project has become an Alphabet business, Waymo. It would appear that the purpose of Waymo is to provide self-driving technology to manufacturers and businesses in the automotive industry rather than to produce its own autonomous car. It describes itself as a ‘self-driving technology company’ and it sees its ‘technology being useful in personal vehicles, ridesharing, logistics or solving the last mile problem for public transport.’ 
A few days after this announcement it revealed its work with Chrysler’s Pacifica Hybrid minivans. Waymo is integrating its self-driving technology in the vehicles, which should be on public roads in 2017. 

Mobileye showcases its system; Apple enters the game

Mobileye – specialists in computer vision systems and mapping – which was previously a partner with Tesla, is now working with Delphi and Audi.  In December 2016 Mobileye released a video of a car using its technology (a cruise control autopilot-enabling EyeQ3 chip) driving without the input of a human. Using eight cameras to monitor its environment it stops at red lights, changes lane when appropriate, brakes in an emergency etc. 
At the Consumer Electronics Show (CES) 2017 Mobileye showed off its Centralized Sensing Localization and Planning (CSLP) automated driving system – a turnkey, fully integrated automated driving solution – which will be ready for production by 2019.
Evidence has also been recently uncovered that Apple also plans to enter the autonomous car market. A five-page letter to the NHTSA in which Steve Kenner, head of product integrity, while expressing approval of the NHTSA’s proposed guidelines, asked for light regulation to allow a level playing field to new and established autonomous car manufacturers. The letter doesn’t say whether Apple is working on its own car or simply working on software and associated technology. Furthermore, on 8th December, more evidence emerged with the discovery of a patent filed by Apple titled ‘Collision Avoidance Of Arbitrary Polygonal Obstacles.’ The patent is for a new collision avoidance system for use on an autonomous vehicle.  

Hacker releases guide and software for Joe Bloggs to do self-drive retrofit 

George Hotz became famous as a hacker who hacked Apple’s iPhone carrier-lock and Sony’s PlayStation 3 computer game console. Now he is CEO of comma.ai; however, he hasn’t given up being subversive. 
In October Hotz had to cancel a product which was designed to turn a normal car into a self-driving semi-autonomous car, due to the US National Highway Traffic Safety Administration’s concerns over the product. Now Hotz and his team of 50 are ‘releasing’ a new product. 
In order to sidestep NHTSA regulation, Hotz is not planning to ‘sell’ anything. Rather his company has posted 3D printing instructions for Comma Neo’s 3D printed housing and posted open-source software on Github, a site which allows software developers to share their own, and contribute to others, software. This means that anyone could download, build and use his product to retrofit cars with comma.ai’s self-driving software.  

VW creates MOIA, a company to focus on ride-shares and car-pooling

At the ‘TechCrunch Disrupt’ event in London in December 2016, Volkswagen announced stand-alone company MOIA. Its focus will be on developing ride-sharing and on-demand mobility services in cooperation with cities and existing transport networks. Volkswagen sees the first project for MOIA in the ride-hailing app market – VW already has a stake in Gett, one of the world’s leading ride hailing providers. It plans to expand the scope of the app, beginning with the growing Moscow market.
The second area MOIA will be working in is car-pooling, or connected commuting. It aims to create a holistic transport solution which optimizes the current transport and road networks. It plans to begin testing the project in 2017. 
On the same day BMW also expressed its interest to get into the ride-hailing app market. It is aiming at the high end premium vehicles area of the market. 

Nissan launches semiautonomous car, hires anthropologists, and announces its autonomous towing system

Nissan’s new Serena car will come with built-in ProPILOT technology, which enables steering, acceleration and braking to be automated on public highways in single lane traffic by the driver flicking a switch on the steering wheel to activate or deactivate the function.  The car will keep distance from that in front, can corner and stay in lane. This is currently only available in Japan, but Nissan says that it plans to build the technology into a new Qashqai for release in Europe in 2017. 
Melissa Cefkin is principal scientist and design anthropologist at Nissan Research Centre in Silicon Valley and is responsible for helping Nissan move its autonomous system into the complex environment of the city. She has been analysing how humans interact with cars, as pedestrians, cyclist and drivers. She has so far discovered that people often interact through eye gaze or gestures. Her discoveries about the varying human-car interactions across cultures could be crucial in cross-cultural implementation of autonomous cars, Nissan believes. 
Nissan recently announced the introduction of its automated intelligent vehicle towing (IVT) system. With the assistance of attached cameras and laser scanners a modified Nissan Leaf autonomously tows trollies carrying vehicles between loading and unloading points at Nissan’s Oppama plant. The system doesn’t use rails or magnetic tape. The cars are connected to a central traffic control system which monitors each car’s location, battery life, driving speed and operational status. Nissan says that it will continue to trial the system at the Oppama plant with the possibility of future implementation at other plants.

Automated railway system plagued by delays

Mining company Rio Tinto has declared that it is continuing its testing of the USD518 million AutoHaul project at Pilbara mine. To date, Rio Tinto has tested its autonomous train system on over 75,000 kilometres of mainlines. Rio Tinto is honest about the delays it is experiencing, stating that due to the delays, expected production from the Pilbara mine will be lower than previous estimates (330-340 million tonnes compared to 350 million tonnes). It expects the Autohaul project to be completed by 2018 – three years later than the original completion estimate of 2015 which it stated in 2012.    

 

Add this: