In the last few weeks Motoring Weekly has introduced two companies who are working in the background of autonomous vehicles: drive.ai and Nauto. Both are definitely not household names like Tesla or Uber, even Waymo is starting to be more well known globally and General Motors has its Cruise Automation division making some noise. Now it is time to introduce you to two more companies currently in the background, however they may become the next big thing: Embark and Starsky Robotics.
Several years ago we highlighted the Otto self driving trucks that were bought by Uber. There appeared to be a bit of smoke and mirrors with these vehicles because when Uber bought Otto, it seemed that the founder of Otto had allegedly stole some documents from Waymo and after a court case, everything was settled financially and the Otto name disappeared. We are now going to see the concept of self driving trucks under two different names and the world wonders whether this is the end of the truck driving job which has sustained countries the world over for many decades.
Embark Trucks have been in business for several years now and got their first truck working back in 2016 packed with technology similar to that seen in the many autonomous car systems being developed. Embark recognised that the easy bit of any truck route could be done without a human – in fact it could be safer without a driver simply because in the US, the long haul routes are primarily long straight multi-lane roadways. This is the boring bit where a driver could easily fall asleep at the wheel if too tired. Embark’s plan was to use a human to get the vehicle to the interstate roads and then take over again when the truck got close to its destination and needed a higher level of skill to navigate city streets.
In 2017, they raised around $15m in capital and partnered with Peterbilt, the manufacturer of the truck they filled with tech the year before. The advantage of that partnership was that a series of test trucks could be built to Embark’s specification – not retrofitting the equipment later, which is always prone to introducing errors or points of failure in cabling etc.
Later in 2017, Embark signed a deal to start delivering products using a small fleet of trucks for Ryder, the trucking company. The trailers were filled with refrigerators from Frigidaire in El Paso, Texas and then they were sent to Southern California – with a human in the tractor unit. Their job was to do the hard work at each end, swapping drivers who then monitored the “robotruck” when it was in control of the driving.
During early 2018, Embark completed a coast-to-coast run from Los Angeles to Jacksonville in Florida taking five days to complete the test – then they went back again! The truck was running on “Level 2” automation meaning that a driver was present to resolve any issues and the company said at the time that a “Level 4” truck (without a human) could do that run in only two days.
Starsky Robotics has been working for a similar amount of time as Embark on their technology which also deals with the monotonous bit of any long-haul truck journey – the interstate runs that are typically on long boring stretches of road. However they have also looked at the issues surrounding loading yards and working on private property as opposed to public roads. By 2017 they had demonstrated a robotruck shuffling trailers around a yard and had even delivered a trailer load of goods for payment.
Like Embark, they also received a chunk of change in capital during 2018 and even had a truck deliver goods after a hurricane, running 68 miles without a driver which was a record at that time. Part of the investment that both companies received that year was due to industry concern that there were more long-haul jobs on offer than people wanting to get into the cabs. So there was a growing demand for their technology.
The difference between Starsky Robotics and all other companies attempting to build autonomous vehicles is that they will not use LIDAR technology. Many companies insist on this laser based technology because it builds a picture around the vehicle. Starsky (like Embark) still want human interaction with the vehicle when in urban areas and they feel that a human can do a better job than a machine in this respect. Remember, a LIDAR system builds a picture and then has another “brain” figure out the what, the where and the how, to navigate. That means that the computer has to have enough data to figure out what to do and when. A human is inherently more successful at this and has other qualities that the computer systems don’t: risk management, empathy and the knowledge of when it is safe to move.
They do use cameras and radar systems that merge the information to ensure that on the easy piece of the journey the system can detect what is going on around them, however in tight or busy areas, a human takes over. One of the critical reasons that Starsky doesn’t want to use LIDAR is range. The range of this technology is currently shorter than the distance it would take to stop a fully laden semi-trailer and that is a concern to the developers. They believe that LIDAR has not developed far enough in reliability, range or ability to help a robotruck.
Earlier this month, Starsky ran a fully autonomous truck at 55 mph on a closed highway in a test that some claim to be the fastest unmanned robotruck yet. Florida is the place to be to see these trucks in action and very soon, they will become commonplace – once they get regulatory approval for longer unmanned journeys.
The US is the perfect place for robotrucks – they had the advantage of building roads in a “modern” environment, i.e. within the last 200 years and with lots of space to build long straight, easy roads to drive on. In the rest of the world, roads were mostly built over 1,000 years or more and were formed by feet, animals and then much later by carriages and motor vehicles. That is why roads in many countries are more fun to drive than the US however not with a semi-trailer!