Receive up-to-the-minute news updates on the hottest topics with NewsHub. Install now.

Driverless cars: Google's Waymo accuses Uber of stealing self-driving "secret recipe"

February 24, 2017 10:17 AM
106 0

Uber acquired self-driving truck firm Otto last year, months after its founder departed Google

Autonomous cars - they've been a staple of science fiction for years, appearing in films like I, Robot, Demolition Man and Minority Report. Thanks to the brightest minds in Silicon Valley, however, they're rapidly becoming science fact.

Google is nearing the final stages of testing for its autonomous car programme, Tesla drivers can enjoy an 'Autopilot' feature for hassle-free motorway driving, and Pittsburgh residents can hail an Uber that drives itself.

But how do driverless cars work? Are they safe? When can we expect to try one out for ourselves? We answer all these questions, and more, below.

22/02/2017: Alphabet's Waymo — Google's self-driving car project — has sued Uber, accusing it of stealing 14,000 files to copy its sensing tech.

Waymo said Uber and the autonomous trucking firm Otto, which Uber bought last year, of stealing plans for Waymo's Lidar, a laser based sensing technology. The complaint filed in California said: "Uber's Lidar technology is actually Waymo's Lidar technology."

Lidar bounces light off objects to "see" where they are — it's not a new technology, but Waymo's version is reportedly a tenth of the price of existing systems, according to Reuters.

"Hundreds of Waymo engineers have spent thousands of hours, and our company has invested millions of dollars to design a highly specialised and unique Lidar system," the company said in a post on Medium. "Waymo engineers have driven down the cost of Lidar dramatically even as we’ve improved the quality and reliability of its performance. The configuration and specifications of our Lidar sensors are unique to Waymo. Misappropriating this technology is akin to stealing a secret recipe from a beverage company."

Waymo said building the technology has taken seven years, while Uber's Otto appears to have done the same over nine months.

The accusations centre on Anthony Levandowski, the founder of self-driving truck firm Otto, which hit the roads last year by shipping a lorry-load of Budweiser beer. In August, Uber bought the company.

Before founding Otto, Levandowski worked on Google's self-driving project, now dubbed Waymo. It alleges that he downloaded 14,000 files before he resigned in January 2016 — including Lidar designs.

"To gain access to Waymo’s design server, Mr. Levandowski searched for and installed specialised software onto his company-issued laptop," the Waymo blog post alleges. "Once inside, he downloaded 9.7 GB of Waymo’s highly confidential files and trade secrets, including blueprints, design files and testing documentation. Then he connected an external drive to the laptop. Mr. Levandowski then wiped and reformatted the laptop in an attempt to erase forensic fingerprints."

Waymo said it spotted the knockoff design via an email slip-up. "Recently, we received an unexpected email," the company said. "One of our suppliers specialising in Lidar components sent us an attachment (apparently inadvertently) of machine drawings of what was purported to be Uber’s Lidar circuit board  —  except its design bore a striking resemblance to Waymo’s unique Lidar design."

Uber said it took the allegations "seriously" and would "review this matter carefully". Waymo is seeking unspecified damages, but asked for Uber to be banned from using the Lidar information.

"Competition in the self-driving space is a good thing; it pushes everyone to develop better, safer and more affordable technology," Waymo said in the blog post. "But we believe that competition should be fueled by innovation in the labs and on the roads, not through unlawful actions."

It's not the first legal battle in the self-driving arena: in January, Tesla sued a former director of its Autopilot automation project for poaching its staff.

22/02/2017: Just over two months to the day after being kicked out of San Francisco, Uber has begun trials of its self-driving fleet in Arizona, this time with the permission of the state's authorities.

Uber's previous test of 16 self-driving cars in California in December was met with opposition from the Department of Motor Vehicles, which claimed the company failed to apply for the correct test permits. Uber always maintained that because a human was always behind the wheel, the cars did not fit the criteria for the $150 permit.

It appears, however, that Arizona has taken an entirely different approach to the testing of autonomous vehicles. In August 2015, Ducey signed an executive order in support of the development of self-driving technology, urging authorities to "undertake any necessary steps to support the testing and operation of self-driving vehicles" on roads within Arizona.

Customers in Tempe are now able to hail a self-driving Volvo XC90 SUV on the Uber service, which will be overseen by two Uber engineers in the front seats, in case a situation arises that requires human intervention.

This signifies a much needed win for Uber given a series of bad press reports over the past few months. Aside from the controversy in California, the technology is far from perfect and has been caught running red lights, as well as narrowly avoiding a collision in San Francisco. The company itself has come under fire this week, with claims of rampant sexism from a former Uber engineer.

Uber has said it plans to expand testing to other cities in Arizona within the next few weeks.

Google has patented a technology that could prevent injuries in the event of an accident.

The company has filed a patent with the US patent office for "adjusting a shear pin to minimize an impact force felt by an object in a collision with a vehicle".

This means its self-driving cars' computers could analyse the amount of force needed to prevent injury or damage to the person, animal or object it's about the crash into.

“Computing devices within a vehicle may determine that an impact with an object cannot be avoided by way of braking, steering, and/or accelerating the vehicle,” the patent explained.

“When this is the case, the computing devices may work to adjust the amount of force necessary to break one or more shear pins which are holding a panel of the vehicle in place. The force of impact on the object which the vehicle collides may be reduced, thereby limiting the severity of injuries and/or damage to the object.”

The computer device inside an autonomous car would receive data that it is about to collide with an object and the shear pin, which determines what will happen to a vehicle when it is subjected to pressure, would move into a position that would cause it to break, meaning the car's panels would slide back and absorb the impact.

“The characteristics of the object with which impact is imminent may be used to determine to adjust the amount of shear force necessary to break the one or more shear pins, in an effort to minimize the impact force felt by the object,” the patent said. “The size of an object may be used to estimate its mass. This in turn may be used to determine the appropriate amount of shear force for the shear pins.”

Google's automated car company, Waymo, hasn't revealed whether the technology will be integrated into its cars or if it will license the technology to other automated car makers.

16/02/2017: Microsoft India has just announced a partnership deal with India's Tata Motors, collaborating on technologies designed to make driving a more personalised experience.

The new partnership will focus around AI, IoT and machine learning technologies, and will provide Tata vehicles with advanced telematics, remote monitoring and sophisticated navigation technologies.

The first vehicle produced as part of the new agreement will be shown off at the Geneva International Motor Show on 7 March.

The company wants to create "a highly personalised, smart and safer driving experience across the digital life of a vehicle owner", according to Tata Motors managing director, Guenter Butschek.

"In the first phase, the advanced offerings will incorporate technologies such as cloud computing, analytics, geo-spatial and mapping and increased human-machine interface, creating a new benchmark in the industry for connected vehicles," he said.

"Using the Internet of Things, artificial intelligence and machine learning technologies, we will provide the vehicle owners here as well as across the world a safe, productive and fun-driving experience," said Microsoft India's president, Anant Maheshwari, adding that Microsoft expects 90% of cars to be completely connected within three years.

13/02/2017: Ford is to invest $1 billion dollars into an AI startup over the next five years, comprised of ex-Google and Uber engineers who have worked on autonomous car projects.

Argo AI will be led by CEO Bryan Salesky, formerly of Google, and Uber's Peter Rander as company COO, both of whom were leaders of their companies' self-driving car teams. The organisation, based at Ford's Pittsburgh headquarters, is expected to have more than 200 members by the end of the year.

This represents the latest push by a car manufacturer into autonomous technology, which has become a highly competitive market over the past few years. Audi, Nissan, Tesla, and even companies traditionally not associated with vehicles, such as graphics processing firm Nvidia, are working on their own technology.

"The next decade will be defined by the automation of the automobile, and autonomous vehicles will have as significant an impact on society as Ford's moving assembly line did 100 years ago," said Ford CEO Mark Fields. "We believe that investing in Argo AI will create significant value for our shareholders by strengthening Ford's leadership in bringing self-driving vehicles to market in the near term, and by creating technology that could be licensed to others in the future."

The deal will see Ford focus on its development of vehicle hardware, including the exterior and interior design of autonomous cars, and the integration of new systems. Argo AI will latch onto Ford's current software development teams and help the company scale up its production.

If all goes to plan, Ford intends to have a fully autonomous, level 4-capable vehicle for the commercial market by 2021. This, according to the Society of Automated Engineers (SAE), is one rank behind "fully autonomous", in which the AI is able to fully control every aspect of the driving and does not rely on the input of a human driver.

"We are at an inflection point in using artificial intelligence in a wide range of applications, and the successful deployment of self-driving cars will fundamentally change how people and goods move," said Salesky, Argo AI's CEO. "We believe this partnership will enable self-driving cars to be commercialised and deployed at scale to extend affordable mobility to all."

Although Argo AI will initially focus on supporting the production of Ford vehicles, the company has stated that future plans could include the licensing out of its technology to other firms wishing to move into the autonomous sector.

Picture credit: Ford newsroom (l-r): Peter Rander, Argo AI COO; Mark Fields, Ford president and CEO; Bryan Salesky, Argo AI CEO; and Raj Nair, Ford executive vice president, Product Development, and chief technical officer.

Drivers who choose to operate self-driving cars will be forced to take out dual insurance policies to cover vehicles in the event they spin out of control while under the supervision of AI.

The Department for Transport is set to unveil plans for the two-in-one policy as it attempts to reduce the legal hurdles preventing self-driving cars from hitting the road in the UK. Who would ultimately be responsible for the cause of a driverless accident has been the cause of significant debate for legal bodies across the world.

The latest plans will introduce policies that cover drivers while they are at the wheel, and then allows responsibility to be handed over to AI while in driverless mode.

Claims made against accidents involving AI will be covered by the insurance company who will recuperate losses from the party responsible, which may include the car manufacturer. The government hopes this will make it easier for victims of crashes involving driverless vehicles to access compensation.

"Automated vehicles have the potential to transform our roads in the future and make them even safer and easier to use, as well as promising new mobility for those who cannot drive," said transport secretary Chris Grayling, in a statement to the Telegraph.

"But we must ensure the public is protected in the event of an incident and this week we are introducing the framework to allow insurance for these new technologies."

Accidents involving driverless cars have raised questions as to who exactly is to blame for a crash. In June last year a driver was killed when his Tesla Model S collided with a tractor trailer after failing to break.

But the government is keen to encourage the growth of the driverless car industry in the UK, an industry analysts predict will be worth almost $42 billion globally by 2025.

Ian Crowder, head of public relations at AA, argued that although improving policies will make insurance claims more streamlined, there is no way to predict how other human drivers will respond to an increase in driverless vehicles.

"In the longer term the number of claims should dramatically fall," said Crowder. "However there are concerns relating to the behaviour of drivers of human-driven cars mixing with autonomous vehicles, which could lead to an increase in collisions."

"For example, frustration at autonomous vehicles sticking to the speed limit; human drivers attempting to fit into smaller spaces between autonomous vehicles etc... which could lead to an increased number of collisions so that could lead in turn, to an increase in premiums overall," added Crowder.

Japanese car manufacturer Nissan is expected to go ahead with its plans to test driverless cars on the streets of London later this month, following successful preliminary tests in Milton Keynes last year.

01/02/2017: A deal with automaker Mercedes-Benz will see its range of vehicles join Uber's ever-expanding fleet of self-driving cars.

The partnership will see Germany's Daimler join forces with the popular ride-hailing service to introduce and operate its own self-driving cars on Uber's network.

"Auto manufacturers like Daimler are crucial to our strategy because Uber has no experience making cars," said Uber chief executive Travis Kalanick. "This became very clear to me after I visited an auto manufacturing plant and saw how much effort goes into designing, testing and building cars."

"That's why instead of building them ourselves, we want to partner with the best auto manufacturers in the world," he added. "We can combine Uber's global ridesharing network with the world-class vehicles of companies like Daimler, so that Uber riders can have a great experience getting around their cities."

Uber has so far declined to offer the financial details of the new deal, however Daimler represents the first company to take advantage of Uber's new policy allowing third-party manufacturers to add their own self-driving cars.

As a result of a $300 million deal with automaker Volvo, vehicles have already been piloted in Pennsylvania and Arizona using self-driving kits developed by Uber.

However unlike previous deals, Uber will not own or manage the new fleet of Mercedes-Benz vehicles, or be responsible for the research and development of the AI led software. Instead, Uber has simply added the Daimler manufactured cars to its self-driving vehicle range.

Tesla semiautonomous cars won't be recalled in the US, despite the technology inside one of its cars resulting in the death of a customer.

Joshua Brown was killed when the Tesla he was driving collided with a lorry in May 2016. However, Tesla had claimed this wasn't because of its Autopilot technology failing, but rather user error, because although the autonomous system can be used as an assistance to the driver, drivers should never take their hands off the wheel.

The US National Highway Traffic Safety Administration (NHTSA), which conducted an investigation into the crash, said data from the car's onboard tracking system revealed Brown had not made any attempt to brake, steer or do anything else to avoid the collision. He would have seen he was going to hit the lorry for seven seconds, which was enough for him to take action and override the Autopilot system, but he decided not to.

"These systems require continual and full attention of the driver," Bryan Thomas from the NHSTA said. "It's not enough to put [information] in the owner's manual and hope drivers are going to read that.

"Manufacturers must anticipate how drivers would use the functionality, and that some don't read the manual."

The NHTSA said it would not take any action against Tesla, and found no faults with the vehicle, but would be monitoring the situation carefully to ensure it does not pose a risk to drivers.

Oxford self-driving car company Oxbotica has announced that the localisation and mapping software used in its prototype vehicles is now ready for general release.

Comprising one part of the company's Selenium OS for autonomous cars, the firm has used its Dub4 software tool in the GATEway pods operating in Greenwich and in the company's own prototype vehicles, which were demonstrated in Milton Keynes last year.

Oxbotica is competing in the same space as Google and Uber, but what sets it apart from other projects is that rather than the expensive laser and GPS systems seen in other cars, Oxbotica's mapping system can rely purely on vehicle-mounted cameras.

"The software uses a single stereo camera mounted onto the car to determine its precise position and orientation in the world," the company explained in a statement. "With this localisation information, the software is then able to create and navigate using vision-based maps in highly unstructured environments without any reliance on GPS or expensive laser-based techniques."

By negating the need for costly LIDAR systems, Dub4 has the potential to make self-driving technology much cheaper to produce, while Oxbotica claims it doesn't require specialised equipment to run; it can be installed and operated on off-the-shelf commodity PC hardware.

"We're excited to be working on deployment of our class-leading Dub4 software solution, as part of our mission to achieve Level 4 autonomy by 2020," said Oxbotica CEO Graeme Smith. "What we've done with Dub4 has never been done before and it represents a seismic shift for the self-driving vehicle industry, enabling a move away from GPS and 3D laser-based approaches. We're paving the way for more affordable and accurate solutions in the industry."

Self-driving cars use a battery of sensors to detect their surroundings, including radar, lasers and camera arrays. The vehicle's onboard computer then uses specialised software to react to this input in real-time, adjusting the car's steering and acceleration to suit the situation.

While some vehicles, such as the Google prototypes which have popularised the idea, are totally autonomous, other cars only automate certain aspects of driving. Tesla's Autopilot feature, for example, features adaptive cruise control, which maintains a safe speed and distance from other vehicles when driving on the motorway.

It also features basic automated steering to ensure you don't accidentally drift into the wrong lane, and although it can perform basic steering maneuvers automatically, it is not intended to replace a human driver altogether.

There is still much debate over whether or not fully autonomous cars should be made widely available. Some say that their safety record is impeccable, while others argue that automobiles are simply too dangerous to be operated by an algorithm.

This debate has been reignited by a number of high-profile Tesla crashes, one of which has been blamed on an Autopilot malfunction. The incidents have raised questions over whether or not self-driving technology is ready for implementation.

On the other hand, Google has emphasised how safe its fleet of self-driving cars are. As part of the testing programme for its autonomous vehicles, the company has posted regular safety logs detailing any crashes the vehicles are involved in, with the vast majority being caused by human drivers in other cars.

Experts remain divided, however, and if self-driving cars are ever likely to be widely commercially available, they will doubtless be heavily regulated, monitored and tested.

In many cases, they already have. Some high-end luxury vehicles already feature 'assisted driving' elements that are close to fully autonomous driving, with Tesla being the obvious example.

Elsewhere, testing of various driverless car projects is underway in the US, with California, Texas, Seattle and more all hosting autonomous vehicle test runs. The UK government has also given the green light for similar trials in Britain, so UK commuters may soon be sharing the roads with autonomous shuttles.

As for when driverless cars will become commercially available, however, that is yet to be seen. As the technology is young and comparatively untested, it is likely that substantial tests will have to be conducted in partnership with government and regulatory bodies before the vehicles are available for public purchase.

There is also a question as to whether or not the average driver will be able to purchase one at all. Even if fully autonomous cars are allowed to be sold, the price of each vehicle is likely to be prohibitively expensive due to the large amount of sophisticated technology that goes into them.

As such, it's likely to be at least a few years before autonomous cars are approved for general road use, and most likely another few years after that before they become a commonplace sight.

As no fully autonomous cars have yet made it to the mass-production phase, no legislation currently exists to govern their public use. However, transport authorities both in the UK and in the US are currently in the process of drawing up new laws or amending old ones to account for driverless vehicles.

In the UK, for example, companies need no permits or authorisation to test autonomous vehicles on public roads. "Real-world testing of automated technologies is possible in the UK today," the government's Pathway to Driverless Cars report reads, "providing a test driver is present and takes responsibility for the safe operation of the vehicle; and that the vehicle can be used compatibly with road traffic law."

The UK also plans to amend the Highway Code to allow vehicles with advanced driver assistance systems to change lanes on the motorway and be parked via remote control. Changes to car insurance would extend it to cover product liabilities for automated vehicles. Insurers would pay out to victims, then reclaim the money from the manufacturers.

The US has also granted testing permits for driverless cars in four states, as well as publishing a 15-point safety checklist for manufacturers of autonomous vehicles in order to better regulate the emerging industry.

Published by the National Highway Transportation Safety Agency, the Federal Automated Vehicles Policy also marks the agency's move to the SAE scale, which measures a vehicle's level of autonomy. The scale runs from zero, where the driver has total control of the vehicles, all the way to five - a totally autonomous car requiring no human input.

Vehicles that measure two or three on the SAE scale - such as the Tesla and other semi-autonomous vehicles - are already road-legal both here and in the US, as the human driver is considered to be in control for the majority of the time.

While many countries have committed to fostering a supportive environment for the development of autonomous cars, concrete legislation governing their use is still likely to be some years away, as the technology itself has not yet been fully finished.

The most well-known companies exploring self-driving cars are Google and Tesla, but there are very few companies connected to the automotive industry that aren't exploring the technology.

In addition to tech companies, however, regular car makers are also getting in on the act. Audi, Ford, BWM, Honda, Daimler, Volvo and more have all been dabbling in some form of autonomous driving project. Luxury passenger vehicles have also been experimenting with using the tech in industrial vehicles such as heavy goods lorries.


Share in social networks:

Comments - 0