Essential Science: Autonomous cars edge to the future

Posted Feb 22, 2021 by Tim Sandle
Autonomous cars are the future, being a question of 'when' rather than 'if', given the levels of investment. It is probable that legislation will one day only allow for self-driving vehicles on the roads. We look at the latest developments.
General Motors: The Cruise AV is designed to operate safely on its own  with no driver  steering whe...
General Motors: The Cruise AV is designed to operate safely on its own, with no driver, steering wheel, pedals or other manual controls.
General Motors Co.
One day we'll all be travelling in autonomous vehicles, at least on the road (autonomous aircraft is a different matter, although this is also something under development). Autonomous vehicles are currently a $50 billion dollar industry, projected to be worth more than $500 billion by 2026. Yet despite the promises of new inventions, the proposes of most vehicles on the road being self-driving remains a little way off. Yet scientific and engineering solutions continue.
A self driving Volvo XC90 outside 737 Harrisonn in San Francisco  Calif.  the unmarked headquarters ...
A self driving Volvo XC90 outside 737 Harrisonn in San Francisco, Calif., the unmarked headquarters of Otto, an autonomous trucking company acquired by Uber in 2016.
Dilu (CC BY-SA 4.0)
This week's Essential Science considers three main developments: how the car 'sees'; how it flows in traffic; and how its power can be improved.
Improving car vision
Central to the autonomous car concept is how a car perceives and processes the external environment. This involves collecting data and making rapid decisions. To process this information, a balance needs to be struck between data processed on a cloud and data processed using edge computing.
File photo: Delphi took their self-driving Audi SQ5 on a 3400 mile road-trip across the US
File photo: Delphi took their self-driving Audi SQ5 on a 3400 mile road-trip across the US
It is also important that the RADAR or LiDAR technology permit the vehicle to navigate safely during times of bad weather, especially where there is poor visibility. To help address this, electrical engineers from University of California - San Diego have developed a way to enhance the imaging capability of existing sensors.
Visualization of section of Sierra Nevada  forest structure developed by the SNAMP  Spatial Team usi...
Visualization of section of Sierra Nevada forest structure developed by the SNAMP Spatial Team using Lidar data.
University of California Cooperative Extension Forestry
The upgrade will enable they vehicle to better predict the shape and size of objects during times of poor visibility. The system has been tested at night and in foggy conditions.
The solution rests on using a combination of light and sound, using multiple sensors, as the following video demonstrates:
To advance the technology, many data sets are needed. To reach the current position, engineers amassed 54,000 radar frames of driving scenes during the day and night in live traffic, and in simulated fog conditions. This provides a starting basis for further investigation.
Of the different navigation systems, Light Detection and Ranging (LiDAR) systems are the most likely to drive the technological future. The current challenge is to find ways to take 3-D imaging systems and reduce their size for vehicle applications. Part of the solution may come from Apple, which is developing very small chips with a view to incorporate LiDAR into a future generation of iPhones in order to improve facial recognition imaging.
The journal Optica has a review of 'shrink the chip' research, in a paper titled "Serpentine optical phased arrays for scalable integrated photonic lidar beam steering."
Improving traffic flow
One of the concerns voiced about autonomous vehicles is that with the 'driver' (or should that be passenger?) unable to speed up or slow down the vehicle, then traffic will become worse. A new study disproves this assumption, provided that autonomous vehicles are working in unison.
University of Cambridge (U.K.) researchers have put together a computer model that shows how a fleet of driverless cars can speed up traffic flow by up to 35 percent. The video below shows how the model would work:
The researchers have proposed a lane-changing algorithm for autonomous cars to work with a fleet of cars that would enable this to happen. For such an approach to be realized, autonomous car manufacturers would need to agree to work together.
All autonomous cars will be electric cars. This means, based on current technology, vehicles charged with lithium ion batteries. This means that irrespective of the sophistication within the vehicle's computer, it will be battery performance defines the vehicle's overall performance. At present, slow charging times and relatively weak power present obstacles.
Charge-discharge diagram for a lithium-air battery
Charge-discharge diagram for a lithium-air battery
For these reasons, considerable research is going into the optimization of vehicle power, both for conventional electric vehicles (like the Tesla) and for future-state autonomous vehicles.
Tesla Model 3 on show at the company s London showroom.
Tesla Model 3 on show at the company's London showroom.
One more recent development (from the Lawrence Berkeley National Laboratory) has been to replace the graphite anodes currently used in electric vehicle batteries with lithium metal anodes (see: "Universal chemomechanical design rules for solid-ion conductors to prevent dendrite formation in lithium metal batteries" published in Nature Communications). These replacement anodes have been shown to extend an electric vehicle's driving range by 30 - 50 percent, there is the issue to overcome into relation to the life expectancy of the battery overall.
File photo: A self-driving car on the road in Mountain View
File photo: A self-driving car on the road in Mountain View
Grendelkhan (CC BY-SA 4.0)
In another development, it has been demonstrated that when charging and discharging battery electrode materials, that high power can be generated by significantly reducing the charging and discharging time without reducing the particle size. By optimizing the ability to create a homogenous electrochemical reaction, better built batteries can be constructed, researchers from Pohang University of Science and Technology have reported.
The research appears in Energy & Environmental Science, and it is titled "Ultrafast kinetics in a phase separating electrode material by forming an intermediate phase without reducing the particle size."
Better for the environment
There has been discussion about electric and autonomous vehicles and environmental impact compared with conventional vehicles. Environmental research finds that, on balanced, electric powered cars lead to lower carbon emissions overall. This is even where electricity generation involves substantial amounts of fossil fuel in the production process. The balanced score card appears in Nature Sustainability ("Net emission reductions from electric cars and heat pumps in 59 world regions over time").
Essential Science
This article forms part of Digital Journal's long running Essential Science series. Each week we take a more detailed look at an important and topical science subject.
A cat called Gizmo  showing all the signs of being sociable.
A cat called Gizmo, showing all the signs of being sociable.
Last week we posed the question of whether a cat's natural hunting instincts can be tamed through good food and interactive play? A new study appears to be confirm that the wilder site of cats can indeed be tamed through diet and exercise.
The week before the focus was with fungi and food. Cereals are a stable foodstuff around the world, yet the crops can be ruined by fungal diseases and this can lead to world hunger. Different studies are being conducted to tackle mycotoxins, and three examples were considered.