As has become customary at this time of year, here’s our Cars of the Future review of notable self-driving developments at CES, “the most influential tech event in the world”.
First, it would be remiss not to point out that self-driving impressively made headlines by not stealing the show! For instance, Cleantechnica ran with the headline: “CES 2023 Shies Away From Autonomous Driving Technology”.
Self-driving premiere
That’s maybe a bit harsh. For starters, ZF gave a world premiere to its next generation Level 4 autonomous shuttle, and announced an important new partnership with Beep.
ZF self-driving shuttle announced at CES 2023
Integrated into ZF’s autonomous driving system is the Virtual Driver software stack, developed in partnership with Oxfordshire-based Oxbotica. It consists of two major parts, the performance path and the safety path. The safety path monitors situations and defines ‘virtual guardrails’, while the performance path enables smooth driving.
“ZF delivers innovative technologies that contribute to sustainable mobility and help decarbonize the world,” said Dr. Holger Klein, CEO of ZF Group. “Today, we have everything to support our customers with holistic vehicle systems based on advanced high-performance controllers, intelligent sensors, smart actuators, connectivity and cloud solutions, and cutting-edge software and functions.”
The agreement with Beep includes plans for “several thousand” Level 4 shuttles in the US. Joe Moye, CEO of Beep, added: “This vehicle will help expand use cases and meet growing customer demand as we continue to pursue our vision of extending mobility equity and reducing carbon emissions with safe, efficient shared autonomous transportation.”
Self-driving AI
Then there was Korean company AIMMO’s announcement of “the world’s first AI-powered Autonomous Driving Data-as-a-Service” – ADaaS – designed “to overcome the industry-wide problem of excessive data collection that has constrained the progression and commercialisation of AV technologies”.
AIMMO AI-powered Autonomous Driving Data-as-a-Service (ADaaS) for self-driving
“Over the years, we have seen a huge amount of anticipation around when we will see autonomous vehicles commercialised, but with standards and regulations ever-changing across the world, it is an extremely complex market to navigate,” said AIMMO CEO SeungTaek Oh. ”We believe that the arrival of AIMMO ADaaS is a game-changer for many companies operating in this space.”
There was the small matter of Honda and Sony teaming up to launch a whole new brand, Afeela. It promises “the car of tomorrow”, with first deliveries scheduled for 2026.
Self-driving tech
Plastic Omnium announced the creation of a new division, OP’n Soft, focused on “mobility solutions that are more electric, more connected, more autonomous and more shared”. There’s probably an acronym for that.
“OP’n Soft will enable Plastic Omnium to offer its customers a unique range of integrated solutions and services, such as merging radar data processing software with lighting technologies,” said CEO Laurent Favre.
Plastic O also showcased a new “smart bumper” featuring embedded antennas to deliver “unequaled sensing faculties”, and announced a partnership with startup Greenerwave “to transform body panels into 4D imaging radar to give autonomous cars supervision”.
Another startup, Exwayz, unveiled SLAM – new generation software offering self-localization accurate to 2cm “to simplify and accelerate 3D LiDAR integration into autonomous systems”.
Exwayz SLAM self-localization for self-driving announced at CES 2023
“We are proud to introduce Exwayz SLAM, aimed at saving years in hard software development to autonomous system manufacturers,” said CEO Hassan Bouchiba. “The reality is that autonomy can only happen with robust, accurate, reliable and truly real-time algorithms, which are the critical lacking elements in currently available solutions.”
Terminator star
Perhaps understandably, these advances were somewhat overshadowed by Terminator star Arnold Schwarzenegger joining BMW CEO Oliver Zipse on-stage to unveil the eye-catching BMW i Vision Dee colour-changing car.
The Dee Movie starring Arnold Schwarzenegger on Youtube
CNBC in America has reported on “an ugly month” for a key electric vehicle and self-driving ETF.
CNBC in America has reported on a key exchange traded fund (ETF) for electric vehicle (EV) and self-driving stocks suffering “an ugly month” in September.
On Friday 30 September, the Global X Autonomous and Electric Vehicles ETF closed 37% off the group’s 52-week high.
CNBC said: “It was the second worst-performing month for the group on a percentage basis on record, behind only March 2020 when the overall stock market saw dramatic declines.”
Self-driving stock
CNBC re EV and self-driving stocks Sep 2022
Global X says the DRIV fund offers high growth potential, noting that: “While global EV registrations increased by more than 40% in 2020, EVs were still less than 5% of new cars sold, highlighting substantial room for further adoption.”
The top 10 holdings (as of 10/04/22) were:
Tesla
Apple
Microsoft
Alphabet
Qualcomm
Toyota
Nvidia
Intel
Pilbara Minerals
Honeywell
At Cars of the Future, we suspect the falls have more to do with the EV side and assisted driving than true self-driving… but rumours of a global recession don’t help!
Transport Select Committee to scrutinise the development and deployment of self-driving vehicles on UK roads.
On 27 June, the Transport Select Committee announced a new inquiry into self-driving vehicles and issued a Call for Evidence.
Chaired by Huw Merriman MP, with a remit to hold Transport Ministers to account and to investigate matters of public concern, the influential cross-party group will scrutinise the development and deployment of self-driving road vehicles.
Transport Select Committee chair Huw Merriman MP to scrutinise self-driving
It follows confirmation that the Transport Bill announced in the recent Queen’s Speech will introduce comprehensive legislation for self-driving vehicles in the UK.
Other heavyweight issues currently before the Transport Select Committee include the integrated rail plan, the national bus strategy and road pricing.
Call for evidence on self-driving
The Call for Evidence on self-driving vehicles reads: “We are particularly interested in receiving written evidence that addresses:
Likely uses, including private cars, public transport and commercial vehicles;
Progress of research and trials in the UK and abroad;
Potential implications for infrastructure, both physical and digital;
The regulatory framework, including legal status and approval and authorisation processes;
Safety and perceptions of safety, including the relationship with other road users such as pedestrians, cyclists and conventionally driven vehicles;
The role of Government and other responsible bodies, such as National Highways and local authorities; and potential effects on patterns of car ownership, vehicle taxation and decarbonisation in the car market.”
The deadline for evidence is Monday 22 August 2022.
Law Commission issues paper on reform options for remote driving on shared roads.
On 24 June, The Law Commission published a 93-page issues paper on reform options for remote driving, inviting feedback from the public.
It notes: “Technology that enables an individual to drive a vehicle from a remote location already exists today – operating in controlled environments such as warehouses, farms and mines.” The new paper considers how the existing legal framework applies to shared roads.
Remote driving
Whether you see it as a bridging technology or a long-term solution, remote driving – where a person outside a vehicle uses connectivity to control it, often from many miles away – will certainly be part of the mix on the road to self-driving.
Indeed, in May, Project Encode demonstrated transfer of control across three states – manual driving, autonomous driving and teleoperation – in live vehicle tests in Oxford and London.
Project Encode remote operation
The issues paper press release highlights four safety challenges:
Connectivity: how can a reliable connection between the remote driver and vehicle be ensured and how can safety risks be mitigated if connectivity is lost?
Situational awareness: how can drivers remain aware of their surroundings through a screen without (for example) the “feel” of acceleration?
Keeping remote drivers alert: how can the risk of fatigue, motion sickness and distraction be overcome?
Cybersecurity: how can unauthorised takeover of vehicles be prevented?
From a law enforcement perspective, tricky questions arise from the possibility that a vehicle on British roads could be remotely driven from abroad. The Law Commission therefore invites views on whether this should be prohibited.
Remote driving terminology
Of course, for anything related to self-driving, there are questions around terminology. In addition to a good old-fashioned driver, and a user-in-charge, we could soon also have an Entity for Remote Driving Operation (ERDO) – a corporate entity rather than an individual that uses and operates vehicles rather than develops or manufactures them.
Nicholas Paines QC, Public Law Commissioner, said: “Remote driving technology is already capable of being used on our roads. We hope our issues paper can contribute to a healthy debate about the appropriate regulation of this technology and what can be done to maximise protection of road users while encouraging innovation.”
The new project – via the Centre for Connected and Autonomous Vehicles (CCAV) and the International Vehicle Standards team at the Department for Transport (DfT) – builds on the recent three-year review of legislation to enable the deployment of automated vehicles (AVs) on British roads.
Views on remote driving can be submitted here until 2 September 2022.
The Law Commission will then draft advice for the UK Government by January 2023.
Self-driving charger robots could remove the need for dedicated EV parking spaces.
Our thanks to Takayuki Yamazaki on Twitter for drawing our attention to ZiGGY, a self-driving electric vehicle (EV) charger.
Aside from looking pretty cool, we think this is exactly the kind of innovation needed to tackle the UK’s notoriously slow rollout of EV charging points.
“In order for the transition from petrol or diesel to electric to be successful, the UK must be able to meet the demand and provide ample charging points for drivers,” she said.
Unfortunately, back in December, The Guardian noted that: “The government has quietly backtracked on proposals to require every shop, office or factory in England to install at least one electric car charger if they have a large car park, prompting criticism by environmental campaigners.
“The original plan required every new and existing non-residential building with parking for 20 cars or more to install a charger. However, the Department for Transport (DfT) has now revealed it will only require chargers be installed in new or refurbished commercial premises amid fears over the cost for businesses.”
The name Ziggy of course brings to mind David Bowie’s fictional alien rockstar, who, according to Wikipedia, “arrives on an Earth that is dying due to a lack of natural resources”.
Self-driving EV charger
This ZiGGY, its LA-based maker EV Safe Charge say, represents “A cost-effective EV charging solution unlike any other. ZiGGY is a robotic mobile EV charging platform that serves all parking spaces, not just a few.”
It goes on to assert that 500 million chargers could be required globally by 2040, up from fewer than six million today, representing nearly $1.6 trillion of cumulative investment in EV charging infrastructure.
ZiGGY self-driving EV charger design features
This nifty robot can be contacted via an app whereby it will secure a parking spot and wait for you.
It will notify you once you’re charged before moving on to the next EV or heading back to base to recharge.
This removes the need for dedicated EV spaces and the addition of video advertising on ZiGGY’s side means there’s a bonus revenue stream as well.
All very clever, but can it play guitar?
On 9 June 2022, Westfield Sports Cars and its self-driving arm, Westfield Autonomous Vehicles, went into administration
In sad news for British motorsport and self-driving fans, on 9 June it emerged that Westfield Sports Cars and its subsidiary Westfield Autonomous Vehicles had both gone into administration.
Founded in 1983 and based in Kingswinford, near Birmingham, Westfield Sports Cars specialised in Lotus Seven inspired kit cars and was often compared to Caterham.
Westfield sports car
Westfield self-driving
Its diversification into self-driving was widely considered an astute move and it gained many plaudits for its involvement in the landmark GATEway Project in London,
In my 2018 article Autonomous Now, which led to the launch of Cars of the Future, I noted: “GATEway is entering its final phase, which will see a fleet of driverless pods providing a shuttle service around a 3.4km route on the Greenwich Peninsula.
Westfield self-driving pod in Greenwich 2018
“In a world first, members of the public are invited to take part in the research, by riding in or engaging with the pods and sharing their opinions.”
Supported by Innovate UK, Westfield went on to run a live commercial trial at Heathrow Airport Terminal 5. The future seemed bright.
Just prior to the Queen’s Speech, on 10 May, Westfield Technology Group CEO Julian Turner was one of the 17 major UK business representatives calling for the Government to announce primary legislation for automated vehicles (AVs).
A slightly bizarre last hurrah – if indeed this is the end – came a few weeks ago when the Westfield Pod featured on Grace’s Amazing Machines on BBC children’s channel CBeebies.
Westfield self-driving pod on CBeebies Grace’s Amazing Machines
For the record, presenter Grace Webb preferred it to the other two contenders – an electric bus and a ride-on lawnmower.
Distressed business listing service, Business Sale, reported that: “In its accounts for the year ending December 31 2021, Westfield Sports Cars reported fixed assets of close to £750,000 and current assets of slightly over £4 million. Less liabilities, the company’s net assets amounted to £1.179 million.
“Westfield Autonomous Vehicles, meanwhile, reported total assets of £1.4 million in its most recent accounts, but ended 2021 with net liabilities of close to £316,000.
“Despite the company’s demise, the assets set to be sold could represent a major opportunity for the right buyer, given their strong offering in the emerging self-driving electric vehicles sector and the niche kit car market.”
Autocar added: “Westfield had built up a solid reputation for creating interesting sports cars majoring on handling and horsepower rather than refinement.
“CEO Julian Turner also planned to push the autonomous pod side of the business, with the aim of turning Westfield into “the Boeing or Airbus of the automotive world”, selling these vehicles to fleet operators.
“Westfield Autonomous Vehicles created the Heathrow Pods that connect the Terminal 5 business parking to the main building and claims to have delivered “more fully autonomous vehicles than anyone else in the UK”.”
Self-driving assets
Mark Bowen of MB Insolvency was appointed as administrator on 9 June, but parent companies Westfield Technology Group and Potenza Enterprises don’t appear to be included.
Mark Bowen told the local media there had already been an “encouraging level of interest shown in the company’s remaining assets” and that MB Insolvency were “liaising with a number of parties at the moment to see if anybody is interested in the assets and possibly trying to resurrect something here.”
On 14 June, the counter on the Westfield Autonomous Vehicles website clocked a not insubstantial “9,983,709 Autonomous Kilometres” and “6,015,384 Passengers Driven”, and still rising.
Surely that should be of interest to someone.
New reports predict self-driving will massively boost the global LiDAR market, with Aeva’s Aeries 4D LiDAR highlighted.
Two new reports have highlighted self-driving as one of the main factors predicted to boost the global LiDAR market to at least US$3.4 billion a year by 2026.
According to Polaris Market Research, the global automotive LiDAR market is anticipated to reach US$4.14bn by 2026, increasing at a Compound Annual Growth Rate (CAGR) of more than 35%.
LiDAR for self-driving
The report summary noted: “The Automotive LiDAR market growth is attributed to the increasing demand of autonomous vehicles for active safety and self-driving. As advanced driver assistance systems (ADAS) and autonomous vehicles are expected to witness growth at significant rates, it is expected to have a direct positive impact on the growth in the Automotive LiDAR market.
“These automated vehicles provide opportunities for a large number of firms to access a range of untapped facts, creating new revenue-generating opportunities, which will boost the market growth.
“The solid-state/flash LiDAR market is expected to grow at a very high pace during the forecast period. Solid state sensor being low-cost, robust, as well as compact in size makes it ideal for potential large-scale production of level 3 and level 4 cars in coming years. Further, mechanical sensors and other sensors also capture decent market share.”
Polaris highlight leading industry players including Scans, Velodyne LIDAR, Quanergy Systems, LeddarTech, First Sensor, Novariant, Delphi, Continental, Robert Bosch and Denso.
A separate report, by Markets And Markets, largely concurs with these findings, projecting that the LiDAR market will grow at a CAGR of 21.6% from 2021 to 2026 to reach US$3.4 billion by 2026.
LiDAR for UAVs
However, it focuses more on unmanned aerial vehicles (UAVs) – drones – and 4D LiDAR specifically.
“The rising adoption of LiDAR systems in UAVs, increasing adoption of LiDAR in engineering and construction applications, use of LiDAR in geographical information systems (GIS) applications, the emergence of 4D LiDAR, and easing of regulations related to the use of commercial drones in different applications are among the factors driving the growth of the LiDAR market,” it says.
“However, safety threats related to UAVs and autonomous cars and the easy availability of low-cost and lightweight photogrammetry systems are restraining the growth of the market.
“The market for 4D LiDAR is projected to grow at the highest CAGR from 2021 to 2026. This growth is attributed to the high adoption of 4D LiDAR in applications such as self-driving cars, robots, and other autonomous systems.
“Apart from automobiles, 4D LiDAR has applications in the architecture, engineering, and construction (AEC) industry, entertainment, and AR/VR. Some of the major companies offering 4D LiDAR are Aeva and TetraVue.”
As well as measuring distance and plotting the position of objects in x, y and z, Aeva’s 4D-LiDAR plots velocity as a fourth dimension.
Aeva CEO Soroush Salehian on self-driving
Soroush Salehian, Co-Founder and CEO at Aeva (formerly of Apple’s Special Projects Group), said: “Bringing Aeva’s next generation 4D LiDAR to the Nvidia Drive platform is a leap forward for OEMs building the next generation of Level 3 and Level 4 autonomous vehicles.
“We believe Aeva’s sensors deliver superior capabilities that allow for autonomy in a broader operational design domain (ODD), and our unique features like Ultra Resolution surpass the sensing and perception capabilities of legacy sensors to help accelerate the realization of safe autonomous driving.”
Gary Hicok, Senior Vice President of Engineering at Nvidia, added: “Aeva delivers a unique advantage for perception in automated vehicles because it leverages per-point instant velocity information to detect and classify objects with higher confidence across longer ranges.
“With Aeva as part of our Drive ecosystem network, we can provide customers access to this next generation of sensing capabilities for safe autonomous driving.”
Big self-driving hardware news as Tesla registers a new radar unit with the US Federal Communications Commission.
Thanks to Angelos Lakrintis on the Linkedin Self Driving Cars group for alerting us to the news that Tesla is apparently doing a major U-turn and re-embracing radar.
The EV specialist famously stopped fitting radar to new cars in May last year, following years of protestations by CEO Elon Musk that self-driving could be best achieved with cameras and silicon neural nets alone.
Self-driving hardware
Last year, The New York Times reported: “Musk has repeatedly instructed the company’s Autopilot team, which works on self-driving car tech, to ditch radar and use only cameras instead.
“The reason for this approach, Musk said in October, is to focus the data that’s being presented to the car’s computer systems.
“Tesla’s camera-based “vision” self-driving tech “became so good,” Musk said, that adding radar data was actually giving the system more information than it needed.”
Musk on self-driving at FT Future of the Car Summit 2022
Indeed, he was still making the point at last month’s FT Future of the Car Summit 2022, saying: “Anyone who’s driven a car for any length of time, once you have some years of experience, the cognitive load on driving a car isn’t that high.
“You’re able to think about other things, listen to music, have a conversation and still drive safely. So, it’s not like matching everything a human does.
“It is matching enough of the silicon neural nets to at least be on a par with the biological neural nets to enable self-driving, and I think we’re quite close to achieving that.”
Well, a week is a long time in politics, they say, and on 7 June Tesla registered a new radar unit with the US Federal Communications Commission (FCC).
Radar for self-driving
It has been widely reported that senior Tesla engineers have long disagreed with Musk on radar, pressing home many of the points made by Clem Robertson, CEO of R4dar Technologies, in our very first Zenzic CAM Creator profile.
“Each technology has its shortcomings,” Robertson said. “GPS is no good in tunnels; the cost of 5G can be prohibitive and coverage is patchy; cameras aren’t much good over 100 metres or in the rain, lidar is susceptible to spoofing or misinterpretation; digital maps struggle with temporary road layouts – but together they create a more resilient system.
“Radar only communicates with itself, so it is cyber-resilient. It works in all weathers. It is reliable up to 250-300m and very good at measuring range and velocity, while the latest generation of radars are getting much better at differentiating between two things side-by-side.”
This latest development suggests that Tesla is now on-board with such thinking.
According to Drive Tesla Canada, the registration allows Tesla to sell vehicles with the new units installed in the US. It speculates that they could form part of the highly anticipated Hardware 4.0 (HW4).
“Tesla currently builds vehicles with HW 3.0, otherwise known as the Full Self-Driving (FSD) computer,” it notes. “It is believed that Tesla will introduce the next-generation computer with the launch of the Cybertruck.
Will Tesla Cybertruck have HW4 for self-driving?
“Whether Tesla will offer existing customers a free upgrade to the new computer, like it did after the introduction of HW3, remains to be seen.”
Given Tesla was previously such a strong advocate for binning radar, it will be interesting to see whether others also back away from the idea.
For instance, Auto Evolution reported in April that Michael Benisch, VP of Engineering at Toyota subsidiary Woven Planet, believes a camera-only approach is possible.
Perhaps tellingly, Toyota itself always remained committed to using multiple sensors, both lidar and radar, on all vehicles offered for sale.
Musk himself calls the motor industry “hyper competitive” and with all major vehicle manufacturers now embracing electric, Tesla’s old USP is no longer unique.
If this U-turn on radar is a sign of a maturing, perhaps more sensible Tesla, its rivals should probably be pleased and worried in equal measure.
Self-driving related highlights from Elon Musk’s keynote conversation at the FT Future of the Car Summit 2022
Following Volkswagen CEO, Herbert Diess, and Volvo Cars CEO, Jim Rowan, on day one of the FT Future of the Car Summit 2022, there was no doubting the biggest draw on day two: an hour-long “keynote conversation” with Tesla CEO, Elon Musk, covering Twitter, Tesla, SpaceX, self-driving and more.
The part live, part digital session was hosted by The Financial Times’ Global Motor Industry Correspondent, Peter Campbell, from The Brewery in London.
Tesla early days
It started with JB Straubel, formerly chief technical officer at Tesla, now Founder and CEO at renewable energy company Redwood Materials, joining Campbell on stage to discuss the origins of Tesla, with Musk contributing via video link.
JB Straubel and Elon Musk at FT Future of the Car Summit 2022
EM: “We got together for lunch and the conversation turned to electric vehicles. JB said I should test drive the tzero prototype from AC Propulsion, that was in 2003. I tried to convince them to commercialise the tzero and, after a while, they said they really did not want to. I said, do you mind if I create a commercial electric sports car?”
JBS: “That’s pretty close to how I remember it. My perspective was us trying to chat with you about this electric hydrogen aeroplane concept, but our conversation completely turned to talking about lithium ion batteries… stringing together large numbers of small lithium ion batteries to potentially have hundreds of miles of range, which seems commonplace today, but in 2003 was absolutely unheard of. You understood that concept better than anyone else.”
They went on to cover the early work on a Lotus Elise chassis with the AC Propulsion drive train.
EM: “An insane nightmare, basically… almost everything about the first design of the Tesla Roadster was wrong. It was just an important thing that needed to happen to move to a sustainable technology future.
“At the time we created Tesla, there were no startups doing electric cars, and the big car companies had really no electric car programmes going. Therefore, unless we tried, they were not going to be created. It wasn’t from a standpoint of thinking, hey, here’s a super lucrative idea.
“There’s an incredibly big graveyard of car startups. They’ve almost all gone bankrupt. You’ve only heard of a tiny number of them, the DeLoreans of the world, but there are hundreds of others.
“The only two American car companies that have not yet gone bankrupt are Ford and Tesla. Tesla almost went bankrupt so many times I lost count. To start a car company is mega pain. It’s the furthest thing from easy money you could possibly imagine.
“The car industry is hyper competitive. Throughout the world, they have entrenched customers, dealers, service, factories, existing expertise – these are veteran armies.”
At this point, Straubel exited, leaving Campbell attempting to elicit answers about the widely rumoured purchase of Twitter. We’ll only cover that very briefly here.
Musk on Twitter
EM: “I think Twitter needs to be much more even handed. It currently has a strong liberal bias. This fails to build trust in the rest of the United States and also perhaps in other parts of the world.”
Peter Campbell and Elon Musk at FT Future of the Car Summit 2022
PC: “Are you planning to let Donald Trump back on?”
EM: “I’ve talked with Jack Dorsey about this. I have the same mind, which is that permanent bans should be extremely rare and really reserved for spam accounts, where there’s just no legitimacy. I do think that it was not correct to ban Donald Trump, I think that was mistake because it is alienating a large part of the country.”
20 million cars a year by 2030
Global media coverage assured, conversation returned to Tesla and the ambition to make 20 million cars a year by 2030.
EM: “There are approximately two billion cars and trucks in the world and for us to really make a dent in sustainable energy, in electrification, I think we need to replace at least 1% of the fleet per year, that’s where the 20 million units comes from. I think we’ve got a good chance of getting there.
“We have an incredible team at Tesla, executing very well and our annual growth rates are faster than for any large manufactured product in the history of Earth. I think the next fastest was the Model T. If that growth rate continues then we will reach 20 million vehicles a year, but we may stumble.”
On raw materials, he continued: “The two main cathode choices are nickel and phosphate. Iron is extremely plentiful and the second biggest element is oxygen. So, I do not see any fundamental scaling constraints. Lithium is also quite common.
“Our goal is to accelerate the advent of sustainable energy. The three pillars of a sustainable energy future are electric transport, stationary battery packs and sustainable energy sources – solar, wind, geothermal and hydro.
“All of Earth can easily be powered by solar and wind, stationary battery packs and electric transport. You could power all of Europe with a section of Spain, all of the United States with a corner of Utah or Texas. Obviously, it would make more sense to spread this out. I invite anyone to do the basic math in megawatts per square kilometre.”
They went on to talk about SpaceX, in particular the Falcon 9 rocket. Classic Musk: “I’m sure we’ll do more than 1,000 times the payload to orbit of all other rockets on Earth combined.”
Then, briefly, China, and the Tesla factory in Shanghai. Finally, with the hour flying by, they got to self-driving.
Musk on self-driving
EM: “I don’t think you need full human level intelligence to drive a car. You don’t need deep conceptual understanding of esoteric concepts or anything like that. Anyone who’s driven a car for any length of time, once you have some years of experience, the cognitive load on driving a car isn’t that high.
“You’re able to think about other things, listen to music, have a conversation and still drive safely. So, it’s not like matching everything a human does. It is matching enough of the silicon neural nets to at least be on a par with the biological neural nets to enable self-driving, and I think we’re quite close to achieving that. Don’t take my word for it, sign up for a beta programme, look at the videos people are posting.
“I’m confident we will get far in excess of the safety level of humans. Ultimately, probably a factor of 10 safer than a human, as measured by the probability of injury.
“It’s around a million people per year dying from automotive accidents, maybe 10 million per year are severely injured. So, with autonomy, the cars driving, or assisted driving right now, but it will be fully autonomous the future, there’s those who didn’t realise they would have crashed, or hit a pedestrian or cyclist.
“It is important to note that we have never said ever that Tesla Autopilot does not require attention. We have always made that extremely clear, repeatedly. You can’t even turn it on without acknowledging that it requires supervision. We remind you of that ad nauseam, so this was not a case of setting expectations that the car can simply drive itself.”
It was Q&A time, so I submitted the question: “Why don’t you change the name of the Full Self-Driving package? It is driver assistance not self-driving. The name causes so much unnecessary criticism.” I didn’t get an answer.
Neil Kennett self-driving question for Elon Musk at FT Future of the Car Summit 2022
To be fair, his hour was nearly done and questions from the audience were stacking up. Classily, he stayed on for a lengthy period of overtime.
Audience Q&A
Here are some of the highlights…
On micromobility: “Scooters are very dangerous. We don’t recommend anyone drive a scooter.”
On building a small car: “There’s some probability that Tesla will do a smaller car.”
On Tesla licencing their products to other OEMs: “They may be interested in licencing Tesla Autopilot full self driving. I think that would save a lot of lives. I would be very open to that.”
On competitors: “VW is doing the most on the electric vehicle front. There will be some very strong companies coming out of China.”
On AI: “We have the best real world AI team in the world.”
On the next big innovation in personal transportation: “Tunnels are underrated, underappreciated. This notion of induced demand is one of the single dumbest notions I’ve ever heard in my entire life. If adding roads just increases traffic, why don’t we delete them? Decrease traffic. I think you’d have uproar. We already have a proof of concept in Las Vegas with a tunnel going from the convention centre to the strip. It’s working really well.”
On super capacitors: “There simply isn’t enough ruthenium. I thought about it quite a lot. Had I continued as a student and done a PhD at Stanford, a theory I had at the time was to use advanced chipmaking equipment to build solid state capacitors.”
On hydrogen: “The number of times I’ve been asked about hydrogen! If you want a means of energy storage, hydrogen is a bad choice. It’s extremely low density, maintaining it in liquid form is incredibly difficult and it does not naturally occur on Earth. So, you either have to split water with electrolysis or crack hydrocarbons. It is the most dumb thing that I could possibly imagine for energy storage.”
And finally, on wanting to die on Mars: “I just said sure, but not on impact! Really, the goal on that front is making life multiplanetary… to preserve life as we know it, not just humans, but also the other animals and plants. So we don’t end up like the dinosaurs.
“You know, there will be natural calamities that occur on Earth – giant meteors and super volcanoes – and we can also do ourselves in, World War III is maybe looking a little bit more probable these days.
“So, I think it’s important for preserving the light of consciousness that we become a multi-planet species and, ultimately, a multi-stellar species.”
You like autonomous and we like self-driving, as US and UK push different descriptors.
The issue of confusing terminology in the, er, self-driving / driverless / autonomous / automated vehicle industry, has raised its ugly head once again, with the US and UK apparently heading in opposite directions.
The oft-quoted Society of Automotive Engineers (SAE) has recently updated its “Levels of Driving Automation”, revealing in itself, reflecting a shift towards “Automated” and away from “Autonomous” in some quarters.
You can see the logic. It isn’t generally good practice to brand something by what it’s not, so “self-driving” certainly has the advantage over “driverless” in that respect.
We’re talking headline descriptors here. Let’s not even get into acronyms… like the fact there’s one letter difference between Advanced Driver Assistance Systems (ADAS) and Automated Driving Systems (ADS) when confusing the two is so dangerous. Grrr.
Self-driving in Korea
Anyway, back to self-driving. We’re not alone in liking it. For example, this month saw the launch of the epically-named Self-driving Robot Alliance in Korea.
Park Jae-young, official for Korean industrial policy, said: “I hope that the Self-driving Robot Alliance will become a place to drive the growth of the domestic self-driving robot market.”
Argument over then, self-driving it is. Not if the largest economy in the world has anything to do with it!
Autonomous in America
Across the pond, Americans are falling out of love with the term “self-driving” and going back to “autonomous”.
You have to question Tesla’s part in all this. It’s a shame, given everything Elon Musk has done for electric cars, that so many hyperbolic headlines are caused by its confusingly-named Full Self-Driving (FSD) package.
It simply isn’t self-driving as the rest of the industry understands it, and it risks drivers misunderstanding what their cars are capable of.
As Forbes reported back in 2020, Tesla has already had to rebrand its Autopilot as Autodrive in Germany. Has its Full Self-Driving (FSD) brand wrecked the self-driving name for everyone else?
Define self-driving
Not in the UK apparently. British Standards group, BSi, with its famous kitemarks, recently updated its connected and automated vehicles (CAV) vocabulary, sponsored by the autonomous-leaning CCAV.
Thanks to lead technical author Nick Reed, of Reed Mobility, BSI Flex 1890 v4.0:2022-03 included some notable additions, not least for “self-driving”.
BSi’s CAV vocabulary includes “self-driving”
Referencing other definitions in the document, BSi defined self-driving as: “Full function of the dynamic driving task (2.1.24), performed by the automated driving system (2.1.7) within its operational design domain (2.1.48)”.
By way of justification, the definition was accompanied by two notes: 1) Although this term is deprecated within SAE J3016 (2021), it is included here because this is the term best understood by the public to mean the definition as stated (which is the same as that for automated driving).
And 2): The Law Commission of England and Wales and Scottish Law Commission joint report (2022) extended this definition with the condition that the vehicle is to drive “safely and legally, even if an individual is not monitoring the driving environment, the vehicle or the way that it drives”.
Peter Stoker, Chief Engineer for Connected and Autonomous Vehicles at Millbrook, nutshelled the issue last year, telling us: “The key to the future of self-driving is education, education, education – for everyone, the public, vehicle manufacturers, the aftermarket… we have to work on the terminology – autonomous, driverless, CAV, CAM – it’s confusing, even to people who know what they’re talking about.”
With marketing teams in the US pushing “autonomous” and the UK and others pushing “self-driving”, the confusion looks set to continue.
