Category Archives: Uncategorized

French, Japanese prime ministers jump on the driverless bandwagon

Posted on by

World politicians are increasingly taking autonomous vehicles seriously. They start to recognize the importance of autonomous technology and the potential to profit from this innovation. French president Francois Hollande just presented a road map to revive French industry by promoting driverless cars and other technologies. Although it does not appear that France is prepared to fund driverless car development directly, they want to help bring about collaborations fostering this innovation. At this point, this may be what is needed most. There are many funds for this kind research available within the EU and venture capitalists are also discovering the high revenue potential in the area of self-driving cars. It is more important to clarify the legal ground rules for testing and operating driverless cars and making sure that these rules allow for fully autonomous operation. Because only fully autonomous operation will unlock the transformative potential of driverless technology and with it the investments needed to bring it about.

France and the other European countries should make clear that the they have the right and the intention to allow the operation of fully autonomous vehicles in their own territory as soon as they are certified to be extremely safe and as long as their operation is technically limited to a clearly specified region within the nation territory. Such a declaration would remove the doubts about the Vienna Convention standing in the way of automotive progress. Removing such legal uncertainties would open the floodgates for investors seeking to gain the pole position in this transformative technology.

France has good potential to accelerate this technology. The nation is open to novel technogies, has a great research base, has leading sensor manufacturers, and – even if the auto makers have invested much effort into autonomous technologies yet – has some well-positioned companies in the space. An example is the company Induct. It has positioned itself on one of the paths towards fully autonomous technology which is most likely to succeed economically: They don’t try to to build the perfect consumer car which can perform autonomously everywhere at every speed; they use autonomous technology to solve the last-mile transportation problem in  a limited and well known area and they initially focus on low-velocity operations which greatly reduces risk and therefore makes commercial success more likely. What they need are testing grounds, i.e. communities, airports, campus grounds where the vehicles can operate and be perfected.They have had problems finding testing grounds in Europe and have now found a more welcoming climate in Singapore – which is probably not what the Frence president had in mind…

Nevertheless, France should not go this path alone. It should at least partner with the other European countries to ensure that Europe remains at the forefront of this technology.

In Japan, prime minister Abe toured the streets around the parliament in Tokyo in three driverless cars (by Nissan, Toyota and Honda) in advance of the upcoming Tokyo Motor Show. Abe has made clear that he wants to advance “auto-pilot” technology as part of this economic policy.Japan has a long tradition of cooperation between government and industry to bring new technologies forward and is also a world leader in robotic technology – a key area that will both profit and play a crucial role in advancing driverless technology.

Overall, it is clear that autonomous technologies are starting to appear on the radar screen of world leaders. There are still too many misconceptions about this technology and the way it will come about. The views are still very auto-industry centric, focused on consumer cars, and fallacious ideas such as the ‘auto-pilot’ analogy still cloud their judgment. But behind all of this, world leaders are beginning to realize that a major innovation is in the making.

 

 

 

Italian low-cost driverless car prototype takes to the roads

Posted on by

On July 13 a low-cost driverless car developed by Italian computer vision expert Alberto Broggi navigated the roads of Parma, Italy in public traffic. The route included rural and urban roads, two freeway segments, traffic lights and roundabouts.

This is a major accomplishment because almost all current autonomous cars use an extremely expensive LIDAR sensor (which costs between 30000 and 70000 USD). Broggi’s car, in contrast, relies on an array of low-cost sensors including stereo cameras and several low-cost laser sensors. All sensors are hidden from view.

Driverless Car BRAiVE, Source: Vislab

The successful experiment (see full length video) highlights the rapid advances in computer vision. It is probably premature to conclude that autonomous cars will be able to rely mainly on computer vision – the weather situation in Parma was sunny (which has its own problems for computer vision) and cameras have difficulties at night. But the experiment is a strong indication that autonomous car technology is advancing rapidly and that different approaches than Google’s may lead to success.

Congratulations to Broggi and his team. We hope that they will be able to secure sufficient funding to take their work from the prototype stage to a full product soon. Besides the vision algorithms much work is needed to optimize the operational capabilities of the car for dealing with typical and rare traffic situations. This requires hundreds of thousands of kilometers of test driving and much fine-tuning that is beyond the capabilities of research institutes. The upcoming European Research Programm for Research and Innovation might be a good starting point for advancing this technology.

Source: Vislab

DARPA virtual robotics challenge: Lessons for driverless cars?

Posted on by

DARPA has just announced the winning teams of their virtual robotics challenge. Each team had to develop software to enable a very advanced humanoid robot to perform disaster-response tasks such as 1) walking to a standard utility vehicle and entering it, 2) driving the vehicle using steering wheel and pedals, 3)  walking on straight, uneven and obstacle-laden surfaces, 4) picking up a fire hose, connecting it and opening the valve. The winning teams will be provided with actual physical versions of the simulated robots and move on to the next level of the Darpa Robotics Challenge.

DARPA used the open source simulation environment Gazebo to create a virtual environment where the software of more than 100 competing teams could be tested. They have created a complete virtual world with a street, houses, obstacles etc. and developed a model of the robot which could then be placed into the virtual world and controlled by the team’s software. This approach has many advantages: Different scenarios can be tested, no physical damage can occur when the software does not behave as expected, the teams can use the simulation to optimize their algorithms etc.

A similar approach could be used to test driverless cars before they are released to the public. Official simulation environments would be needed with standard interfaces for the driverless car software. This would have to include standardized models which describe the behaviour of the car’s sensors (and translate the state of the simulation environment into sensor readings).

A  standard simulation environment for driverless cars would enable a testing body to verify that a software-based driverless car behaves as expected in many difficult cases. It would allow quantitative claims about the performance of driverless car software and could be used to compare different driverless car systems. Driverless car companies could use such test environments to prove that their software has been properly and carefully developed.

Of course test cases would have to be challenging and representative of the actual scenarios which a driverless car will encounter. This requires significant effort. It is likely over time the number of test scenarios which a driverless car software will have to pass will increase. In the event of driverless car accidents, the testing agency could rapidly analyze the causal chain for the accident and then build additional simulation test cases which the updated driverless car software would have to pass.

As we have argued in our white paper three years ago, it is extremely important to develop a public common simulation platform for driverless cars now!

McKinsey report on disruptive technologies fails to see full implications of driverless cars

Posted on by

It seems to be hard for some auto-industry insiders (including their consultants) to grasp the true nature of the coming changes in mobility. Strategy consultants McKinsey’s Global Institute just published a report on disruptive technologies which correctly envisions significant reductions in accidents and fuel consumption caused by autonomous car technology but the authors don’t really consider the fundamental change in personal mobility – switching from car ownership towards using driverless car sharing – which will be caused by driverless cars!

Driverless cars have the potential of slashing the total number of cars by a factor of 5 to 10 (see the recent Earth Institute study) and therefore will completely and thereby fundamentally transform the auto industry. McKinsey should have given much more thought to this potential disruption as this needs to be factored into the decisions of their automotive industry clients even today (they do have a one-liner mentioning that driverless cars could increase car-sharing but they fail to see the implications). This transformation will occur naturally because of the enormous cost advantages of driverless car sharing over traditional car ownership.

Nevertheless the McKinsey report gets it right when they point out that there are significant first-mover advantages in autonomous mobility and they also have interesting thoughts about the impact on trucking (50% autonomous driving possible by 2025).

The authors may have been involved with intelligent transport systems because they propose that countries should invest in sensors to be embedded in the roads to ‘provide precise positioning information and unambiguous information about speed limits’. This approach is not up-to-date any more. A more adequate and much less expensive delivery mechanism for such information would be the mobile internet: cars could eventually be required to download official maps with such information via the mobile internet.

Developments in Road Transport Automation – EU Workshop

Posted on by

On March 7 another pan-European workshop on Automation in Road Transport took place in Brussels. Europe is and has been very active in automated and autonomous cars with numerous projects having received large amounts of funding over the last 20 years.

Several projects have concluded in the last years, including SARTRE (road trains of networked cars) and CityMobil. Work has begun on CityMobil2 (no website yet) which will implement autonomous transportation system demonstrators in five European cities and which also will work on the legal aspects and industrial potential of this technology.

A key topic at the workshop was the 1968 Geneva Convention on Road Traffic which limits the adoption of autonomous cars because it specifies that a driver must be in control at all times (paragraphs 8.1,8.5 and 13.1). The US and Japan did not sign the convention and therefore may have a crucial head start in the adoption of this technology (we can already see this happening in California and Nevada).

More details about the workshop can be found here.

 

 

 

 

 

 

PwC predicts collapse of car sales because of self-driving cars

Posted on by

PricewaterhouseCoopers – the world’s largest professional services firm – has just released an analyst note about the effects of autonomous cars on the auto industry. While the report is extremely positive about the technology (predicting a reduction of traffic accidents by a factor of 10) it cautions that the fleet of vehicles in the  United States may collapse from 245 million to just 2.4 million. This is a reduction by the factor of 100 and significantly higher than the factor of 10 provided in a recent study by  the Earth Institute which we highly recommend.

It is encouraging that the major consulting firms and think tanks are beginning to take note of the tectonic shifts which will occur in the auto industry within a few years – and which we have emphasized for the last 3 years. The study contrasts with a recent report by KPMG on “Self-driving cars – the next revolution“. While KPMG’s analysts briefly mentioned on-demand mobility services (autonomous car sharing), they failed to see its disruptive potential.

It is time for the auto industry to seriously plan for this future. Contact us – we can help!

 

 

Thrun sounds cautious note on self-driving cars

Posted on by

We just came across an interesting interview by Sebastian Thrun, head of Google X and mastermind behind Google’s Driverless Car. In the interview conducted by Charlie Rose in late April he talks about Google’s project Glass, Udacity and driverless cars. He sounds a cautious note on the safety of autonomous cars. While the technology is already quite safe, he is still concerned about its capability of driving millions and millions of miles without error. He suggests that current driverless cars have not quite reached the perfomance level of an attentive human driver. In another interview with WIRED he also discusses the problem hinting, that the right combination between human and computer intelligence may need to be found to ensure maximum safety.
Are we right to conclude that Google’s driverless car team is finding it difficult to reach the intended safety level for their cars? From a statistical perspective it is very hard to prove that an autonomous vehicle can perform flawlessly for millions of miles. Just driving a few million miles in test mode is not enough.
So far Google has always emphasized that their driverless cars drive on known routes for which detailed navigation and localization data is available. Relying on stored information about a route, however, can be a major source of error. Therefore Google must also be working hard to run their cars without prerecorded mapping data. It would be interesting to know their progress in this area.
An additional approach for verifying the safety of their cars would be to develop a full-fledged simulator built around a physics engine which would be connected to the sensor and actuator interfaces of Google’s driverless car (for an early sketch of this approach, see our Innovation Brief (2010)). The driverless car control unit would receive sensor inputs generated by the simulator. These sensor inputs would be updated according to the simulated movement of the car depending on the signals received by the actuators. Building such a simulator would be a major challenge because of the large amount of sensor data which it would need to generate. But Google’s team has already collected much of the real-world data needed to populate the simulator.
The advantage of the simulator would be that the driverless car could be subjected to much more rigorous testing and that it would be much easier to detect and precisely test borderline situations. Invalid or problematic sensor data could be sent to the control unit, slippery surfaces, fog, rain and snow could be simulated by the physics engine. Unexpected behavior by vehicles, cyclists etc. could also be simulated and tested.

Wall Street Journal considers driverless future

Posted on by

An opinion piece in the WSJ on July 17 criticized investment in the proposed bullet train between Los Angeles and San Francisco. It then advocated preparing the road infrastructure for the advent of driverless cars: The number of lanes should be increased because driverless cars can cope with narrower lanes. Trucks should be limited to one or two wider lanes (ideally separate from car lanes). Traffic flow control systems would have to be installed to optimize traffic flow – both on the highway and in the city where stop lights often increase congestion.

The author points out interesting issues but fails to recognize the decentralized nature of the driverless revolution. These cars do not require modifications to the existing infrastructure. They will drive where a human can drive. But their behavior and economics differ. They will naturally improve the flow of traffic. Driverless cars increase the peak capacity of the existing infrastructure because they can safely keep shorter distances to other cars. Even a small percentage of driverless cars mixed into regular traffic can reduce congestion by adopting an optimal speed or by simply keeping side-by-side with another (automated?) car in two-lane pile-ups effectively preventing other cars from lane-hopping. The cost-structure of driverless car sharing and the need for demand management by mobility providers are also important factors that will reduce peak loads on the traffic infrastructure.

There is no need for driverless road infrastructure investment. Governments would be well advised, however, to invest directly into advancing driverless technology. Fewer traffic accidents would save hundreds of millions of dollars every year. Better road utilization exhibited by driverless cars will also greatly lower the costs for building and maintaining the road infrastructure.

 

 

 

Perfecting driverless cars on the race track

Posted on by

We just came across a talk by Chris Gerdes at the recent TEDxStanford conference. He is busy developing autonomous race cars. Driving these cars autonomously at their limit at high speed, difficult tracks and on slippery surfaces greatly helps improve the algorithms and could also be important to increase acceptance of this technology.
The team has is also carefully analyzing the behavior of professional race drivers to learn about optimal car handling in critical situations. They have gone so far as to equip race drivers with sensors to continuously measure their cognitive load while on the race track. Gerdes hopes to replicate some of the maneuvers that race drivers handle almost instinctively and with very little cognitive load.