Waymo opens-up data treasure trove for autonomous vehicles

Waymo has pulled back the curtain on valuable datasets to help researchers better hone self-driving algorithms.

While this is a nice gesture from the team, we suspect the lid will be kept shut on further datasets unless the idea becomes more mainstream. Data is king in the world of autonomous vehicles and this could prove to be a valuable bonanza for researchers and application developers throughout the world.

Waymo has said the datasets are not available for commercial use, though researchers in commercial organizations are free to access the data for their own development purposes.

“When it comes to research in machine learning, having access to data can turn an idea into a real innovation,” the team said in a Medium post.

“This data has the potential to help researchers make advances in 2D and 3D perception, and progress on areas such as domain adaptation, scene understanding and behaviour prediction. We hope that the research community will generate more exciting directions with our data that will not only help to make self-driving vehicles more capable, but also impact other related fields and applications, such as computer vision and robotics.”

When you look at the development of autonomous vehicles, nothing is more valuable than the right data, and those who collect it are usually very protective. Part of the reason for this is the effort which must be exerted to collect it, with companies like Waymo clocking up millions of miles on the road.

This release contains data from 1,000 driving segments, each capturing 20 seconds of continuous driving, corresponding to 200,000 frames at 10 Hz per sensor. Each segment contains sensor data from five high-resolution Waymo lidars and five front-and-side-facing cameras, offering a 360° view, as well as a total of 12 million 3D labels and 1.2 million 2D labels.

Such data would allow researchers to train models to track and predict the behaviour of other road users, as well as simulate certain situations to find the most appropriate outcome. The dataset covers various environments, from dense urban to suburban landscapes, as well as during day and night, at dawn and dusk, in sunshine and rain.

What is worth noting, as while this is the largest release of data for autonomous vehicles, it is not the first. Lyft released data last month, and Argo AI did so the month before.

The more data which is released to researchers, the quicker the autonomous dream can be realised, and the safer the final product will actually be. It does technically lessen the commercial edge of these organizations, but the final goal of getting autonomous vehicles on the road sooner rather than later seems to be more valuable.

Waymo learning from Darwin for autonomous driving

Google subsidiary Waymo has been working alongside its AI cousin DeepMind to develop a technique called ‘Population Based Training’, based on Darwin’s concepts of evolution.

Although we plan on dumbing down the explanation here, we do also hope to remain true to the work Google’s autonomous driving subsidiary Waymo and AI unit DeepMind are doing to advance self-driving algorithms. It’s an incredibly complicated field, but it does seem like the duo is making progress.

“Training an individual neural net has traditionally required weeks of fine-tuning and experimentation, as well as enormous amounts of computational power,” a blog post stated. “Now, Waymo, in a research collaboration with DeepMind, has taken inspiration from Darwin’s insights into evolution to make this training more effective and efficient.”

The easy part of autonomous driving is almost finished. Sensors are almost up-to scratch and prices will come down quickly when economies of scale kicks in, while the chip giants are making progress also. The trickiest part of the equation is the ‘intelligence’ aspect, the AI components which control all of the decisions.

The simplest way to explain training algorithms is through trial and error. The algorithm performs a task, then grades its performance depending on the outcome. Depending on the ‘grades’ the algorithm will adjust how it performs the task to create a more likely positive outcome.

The challenge which engineers and data scientists face is how much freedom the algorithms are given to adjust with each trial. Too little variance and the fine-tuning takes too long, too much and the results vary wildly. Most of the time, engineers will monitor the tests, manually culling the poorest performing results.

The new approach from Waymo and DeepMind is an interesting one. Population Based Training starts with multiple different tests, before the poorest performing ones are culled from the population. Out of the ‘survivors’, copies are made with slightly mutated hyperparameters. This process goes on and on until the algorithms become more reliable, resilient and safe.

It might sound like a simple solution, but not many companies like Waymo are fortunate to have such smarts as DeepMind living in the same corporate family. Its almost unfair, and we’ve quite surprised its taken so long for Waymo to cosy up to its smarter cousin.

Waymo bags 62k minivans for self-driving service

Waymo has expanded its partnership with Fiat Chrysler agreeing to add up to 62,000 Chrysler Pacifica Hybrid minivans to its self-driving transportation service.

The service, which Waymo plans to launch at the end of the year, will allow the public to use Waymo’s app to request a vehicle will make use of the vehicles, though there is also the potential to make the vehicle available to retail customers. The partnership itself dates back to 2016.

“Waymo’s goal from day one has been to build the world’s most experienced driver and give people access to self-driving technology that will make our roads safer,” said John Krafcik, CEO of Waymo. “We’re excited to deepen our relationship with FCA that will support the launch of our driverless service, and explore future products that support Waymo’s mission.”

“FCA is committed to bringing self-driving technology to our customers in a manner that is safe, efficient and realistic,” said Sergio Marchionne, CEO of Fiat Chrysler. “Strategic partnerships, such as the one we have with Waymo, will help to drive innovative technology to the forefront.”

The Waymo/Google vision is to ‘help people get from A to B at the push of a button’, but is also another excellent example of the company’s ability to persist with investment, keeping an eye on the long-term horizon. Just like the Maps offering and investments in artificial intelligence, there is no immediate gratification in autonomous vehicles, but the team is persisting.

Of course, for those who would not be seen dead in a minivan, the Waymo team has got you covered. Back in March Waymo announced another partnership with Jaguar Land Rover to develop a premium electric fully self-driving vehicle for the transportation service. Up to 20,000 I-PACEs will be added to the fleet, which we can only imagine is for the type of person who deems themselves too important to take the bog-standard Uber service.

Google fuels autonomous vehicle publicity campaign

Waymo – the new name for Google’s autonomous vehicle efforts – is one of a number of organizations behind the ‘Let’s Talk Self-Driving’ initiative.

The purpose of what is described as a public education campaign seems to be to reassure the sceptical mass market that being driven around by a robot is a great idea. The other participants are Mothers Against Drunk Driving, the National Safety Council, the Foundation for Blind Children, the East Valley Partnership, and the Foundation for Senior Living.

Waymo has been running a, ‘early rider’ programme in Phoenix, Arizona, in which regular families have been given access to a self-driving car, with the apparent aim of proving that there’s nothing to worry about. Smiling, wholesome families are shown having their quality of life improved by not having to do the driving themselves (see video below).

That seems to have been the first phase in a protracted attempt to persuade understandably nervous punters that self-driving is safe as houses. “When 94% of road crashes today involve human error, self-driving cars promise a future where anyone can ride with a driver that never gets drunk, tired, or distracted,” says the Waymo announcement.

The campaign acknowledges that it is still early days, and we’re presumably still a long way from the technological, regulatory and social developments needed to make self-driving a mainstream thing, but the movement seems to be ramping up. Of course another major concern about automation is that it makes people redundant, so all this talk of how much more fallible humans are than machines has the potential to back-fire.


Google’s self-driving cars are almost here, as long as you’re not on the left

Google spin-off Waymo is on the verge of cracking the autonomous vehicles conundrum, just make sure wherever you want to go is on the right hand side.

Many are predicting the rise of the autonomous car over the next ten years, with manufacturers giving various timelines around a 2020 deadline, but these predictions could be blown away if Waymo is anywhere near accurate. According to The Information, Waymo could launch its own ridesharing service, not requiring back-up drivers, before the end of the year in Chandler, a suburb of Phoenix.

Autonomous driving might not have gained as much attention as it might deserve so far, but this is primarily due to the timeframes involved. Other areas have quite rightly stolen the limelight because there is a more realistic opportunity to see them in the real world in the nearer future; should these reports prove to be anywhere near true, we could be about to witness a mass panic in the industry, as competitors scramble to get their own versions onto the road.

But there is one minor issue for the moment; the cars struggle to turn left in certain places. A couple of areas which need improvement would certainly be expected, but this is a pretty big issue. A car can only turn two directions; Waymo is missing 50% of this equation. Okay, we’re blowing this a bit out of proportion, as the cars can turn left sometimes. But realistically, your correspondent would want a self-driving car which could turn left all the time. We don’t feel this is a diva-style request.

The problem here lies in a type of junction which the whizz kids in Silicon Valley don’t have on their own roads. It’s a left hand merge, onto a five lane road, where drivers have to jockey for position without causing a hazard for oncoming traffic, but also making sure the manoeuvre is done quickly enough to avoid frustration for following drivers.

“While this type of traffic signal is rare in our hometown of Mountain View, CA, it’s become common sight at intersections across Metro Phoenix, where we recently launched our early rider program,” said James Stout, Lead Software Engineer at Waymo.

“Just like for humans, the key to learning is practise. That’s where our simulator comes in. Waymo’s simulator is a realistic virtual world where we can recreate every real-world mile we’ve drive. Each day, as many as 25,000 Waymo self-driving cars can drive up to 8 million miles in simulation, testing new skills and refining old ones.”

It’s all about practise here. The team will put the vehicles through millions of miles of simulations, with thousands of different variations of conditions and hazards. Just like a human, the more the AI practises dealing with these situations, the more competent it will become. The Waymo team claim its cars drove more than 2.5 billion virtual miles in 2016. Admittedly, this is a virtual environment, but each journey can be customised to be a bit more eventful than your average trip to the shops to buy some milk. It’s just those dastardly left turns…

Even with such a challenge, the team seems to be miles ahead of the competition. Waymo (and Google for that matter) has been relatively quiet in recent months, but actions speak louder than words. Should the self-driving cars hit the roads before the end of the year, Waymo would put itself years ahead of the competition.

That said, we don’t plan to head to Phoenix any time in the near future, but if we do, we might wait for Stout and his cronies to master the left hand turn before using one of their vehicles. We’re just a bit funny like that.