Freight expectations of self-driving vehicles

Jon Lawson
Sam Clark offers an insight into the future of autonomy. It might not be what you think…

The advent of autonomous travel will have far-reaching effects on nearly all aspects of modern transport. From cars to trucks, planes to boats, both long established OEMs and new start-ups are competing to create the perfect autonomous solution. This push has given rise to a range of technologies created to facilitate driving autonomy, with OEMs experimenting with different combinations to move up the five levels of autonomous driving.

Five levels of vehicle autonomy

In an effort to categorise developments in autonomous driving, the Society of Automotive Engineers developed a five-level scale outlining driver input. Level zero is no automation, with the driver in complete control of the vehicle. Level one features basic driver assistance such as cruise control. Level two is occasional self driving, with the vehicle able to change speed and lane in certain environments such as motorways. Level three is occasional self driving where the vehicle is in full control monitoring both road and traffic. Level four is full self driving, with the vehicle in control on an entire trip. And finally, the holy grail, level five is where there is no driver in the vehicle at all.

To date, the best autonomous vehicles are at the level two mark, with some OEMs touching into level three, such as Tesla with its Autopilot feature. Predictions as to when we’ll see true level four and five autonomy vary wildly. Ford’s former CEO Mark Fields predicted in 2017 it would have level four cars on the road by 2021; BMW and Daimler claim it’ll be 2024. Regardless of when it happens, the OEMs are all using various combinations of the same technology to get them there.

Maturing self-driving vehicle technology

Autonomous vehicles depend on one or more of a myriad of technologies to visualise the world around them: light detection and ranging (LIDAR), radar, GPS, ultrasonic sensors and cameras are all leveraged to one degree or another by competing companies to achieve autonomy. LIDAR is frequently the go-to for early autonomous start-ups, and is maturing quickly. Based around a narrow infrared laser shooting short pulses at an extreme rate, LIDAR measures the return speed of each pulse to give the vehicle a sense of depth when determining what objects are in it’s view. Laser stability is achieved not by turning itself off and on, but instead by using a Mach-Zehnder modulator to interrupt one of two channels of light, effectively cancelling out the amplitude of the optical wave to create the pulses cast forward. Anyone familiar with early autonomous vehicle prototypes will know how to identify them by the small dome LIDAR units mounted on the vehicle’s roof. These domes host a spinning LIDAR array measuring and drawing the environment around them. Next-generation LIDAR systems are now doing away with roof mounted domes, instead opting for groups of integrated photonic lasers reduced in size and mounted in the vehicle’s headlights. These arrays will allow steerable beams, further increasing the resolution of the imaging platform.

Many of the largest OEMs are pushing LIDAR, such as Volvo partnering with US tech company Luminar; Audi with Bosch and Ford developing its own in-house system. One notable exception is Tesla, with its forever-vocal CEO claiming last year: “LIDAR is a fool’s errand, anyone relying on it is doomed.” Elon Musk views LIDAR as a crutch, and instead is a proponent of vision-based sensors combined with a neural network to decipher roads and objects. Tesla’s vehicles currently feature eight cameras, 12 ultrasonic sensors and a front-facing radar, though questions remain as to whether this will be enough to take its vehicles to level five autonomy on their own. Part of Musk’s justification for ignoring LIDAR is price, but as Sam Pfeifle asked on the Spar 3D website, at US$250 a unit for a US$37,000 car, is it that big a problem? To date, Tesla is sticking to its non-LIDAR approach, but it’s likely the firm will add LIDAR further down the line, if for no other reason than that it will add an additional layer of redundancy.

Regardless of which combination of sensors OEMs opt for, all are locked in a race to bring truly autonomous vehicles to the driving public first. It’s worth noting though, that while the buzz is frequently focused around self-driving cars, it could be another vehicle that beats the car to the fully autonomous punch.

Self-driving vehicles and the freight industry

The trucking industry is undergoing its own autonomous revolution, with start-ups and established OEMs investing no small sums in a bid to dominate the sector. Daimler announced back in January 2019 that it was investing a whopping US$570m into autonomous truck technology. September 2019 saw the fruits of its labour rollout, with testing starting in Virgina, USA of level four autonomous trucks. Amongst these heavy-hitters, smaller start-ups have emerged vying for supremacy in an increasingly crowded sector. TU Simple and Embark are examples of Silicon Valley-based start-ups that are combining the full suite of autonomous sensors to enable level four autonomy.

Benefits of self-driving vehicles in the freight industry

Autonomy in the consumer car world serves to free up time for the passenger and ultimately improve safety. In the world of freight, time is money, and if mere mortals can be replaced, or at least augmented with autonomous systems, then the dollar value of each delivery can be driven down and the volume of deliveries pushed up. Embark’s CEO Alex Rodrigues predicts autonomous trucks will be able to travel 2.5 times further in a single day than their non-autonomous counterparts as there will be no need to adhere to the 10-hour per day driving ceiling regulated for human drivers. Fuelling this push for autonomy is the ever-growing demand for more capacity as e-commerce providers such as Alibaba and Amazon continue to grow.

The logistical model of businesses such as Amazon sees goods moving between large distribution hubs connected via motorways, before smaller, light goods vans handle final delivery. These hubs tend to be connected via long motorways, the ideal environment for autonomous vehicles thanks to predictable traffic, steady cruising speeds and typically well-maintained roads. Beyond single autonomous vehicles, OEMs such as MAN are trialling multiple truck platooning systems. In these autonomous convoys, one or more trucks follow a lead vehicle maintaining a fixed speed and distance, with all the trucks in the convoy communicating with one another in real time. MAN predicts trucks moving in this fashion will benefit from slipstreaming, reducing C02 emissions and fuel consumption by up to 10% while shortening the distance between vehicles easing traffic and congestion.

One interesting hypothesis is that when level four and five autonomous trucking is achieved, we may see less trucks on our roads rather than more. These levels of autonomy would allow drivers to sleep while the truck drives, increasing the number of routes running through the small hours of the night. It’s also predicted that running on quieter night time roads will likely see a reduction in vehicle collisions, and any improvement in safety can only be a good thing.

A final factor pushing the need for autonomy in trucking is a chronic shortage of drivers. It’s currently estimated that in Europe alone there is a shortfall of 150,000 drivers, with 30,000 each year leaving the profession and only 2,000 coming in. Autonomous and platoon driven trucks could relieve the perceived tedium of long distance driving enticing more drivers to enter the profession. With all of these advances there is understandably a fear that autonomy could remove the need for the driver completely, however it’s unlikely this will be the case. The first and last mile of travelled routes will likely always need driver intervention, and having a person in-vehicle to handle unforeseen events such as a breakdown or road blockages will be crucial when technology can no longer aid.

The road ahead for freight self-driving vehicles

In an industry that has seen little change for nearly a century the pace of the autonomous revolution can seem as daunting as it is exciting. For some truckers it spells the end of an era, for others a change that can’t come soon enough. For the start-ups and OEMs racing to get to level five autonomy there will be huge amounts of money to be made as fleets are augmented or replaced with the latest technology.

For the haulage companies and businesses that depend on them there is a once in a generation chance to slash costs, increase capacity and optimise logistic supply chains. At the same time, governments will play their part in either hindering or supporting the push for autonomy, with legislation adapting as long distance trials are successfully completed. And finally, there is us: the public. How will we feel when passing a 40 tonne truck at 70mph with no driver behind the wheel? All this still remains to be seen, but whatever happens, change is coming and in the world of automotive technology, it’s never been more exciting.

Sam Clark is CEO of Conjure

Recent Issues