Why a challenge like this?
This is a unique test that follows other events which were similar in nature: vehicles have been demonstrated to be able to move in unknown and extreme environments like the Mojave Desert in Nevada/California back in 2005 during the DARPA Grand Challenge; vehicles showed excellent performance when driving downtown in 2007 during the DARPA Urban Challenge. Autonomous vehicles demonstrated great potential, but so far all these tests were conducted in fairly simplified scenarios (with no other traffic in the first case, and with well structured rules in the latter). Other groups are now trying different extreme tests, such as the equipment of a vehicle to drive a rally race. With this event we would like to test and stress our technology on a route that is long (13000 km) and extreme (including all sort of traffic, weather conditions, road infrastructures, and even off-road) to really assess the performance of our systems.
Has anything like this been tried before?
No one has ever tried anything like this. It’s the first time in history that vehicles with no driver and without using oil-based propulsion travel on an intercontinental route. If everything will go as planned, this will be a huge milestone in the history of mobility and robotics.
Who ideated this event?
VisLab did. The idea dates back to 2007, but it was only in January 2009 that we got in contact with Overland and we started planning also the logistics. In 2007 VisLab proposed a similar project to the Italian Ministry but is still waiting for an answer.
Travelling from Italy to China with no driver? It’s impossible: is there a trick?
There’s no trick: there are rules, instead. Since there are no maps of some of the areas in which the vehicles have to move, it would be impossible to ask the vehicles to determine the route. Therefore the convoy is led by a vehicle in which operators take control every time a decision on the road has to be taken. This is the leader vehicle; it is manned all the time, although it runs autonomously most of the time. It is used to test innovative systems and conduct experiments. The leader also broadcasts its GPS position via radio. The second vehicle, i.e. the follower, receives the leader’s GPS position; this vehicle follows the leader’s route defined by very rough GPS waypoints (remember hat DGPS is not available in many areas) and refines its trajectory via local sensing. If a map would be available, there would be no need for a leader. The second vehicle is readily exploitable as a fully autonomous vehicle.
Why are you using such small electric vehicles for a huge trip like this?
Small vehicles are a must for inner city applications. Electric propulsion is a key to sustainable mobility. Pulling the two factors together results in a straightforward choice for this kind of demonstration. Indeed, a 13000 km extreme trip like VIAC is not a good representation of inner city mobility, but if the vehicles survive, we’ll be sure that they will survive in future urban applications too. In other words, we are pushing this technology to check its limits.
The vehicles look strange: couldn’t you integrate the various systems better?
Yes, it would have been possible to integrate the sensors in a better way, but we didn’t do it for specific reasons:
- we are travelling in an extreme environment and each system must be reached easily for quick maintenance. Therfore we preferred ease of use rather than a nice integration
- The whole event is a show; therefore we choose to make all sensors and systems clearly visible and sometimes we even indulged towards a more ‘technological’ look rather than hiding all sensors with a nice integration
Other vehicles, like BRAiVE (see www.braive.vislab.it) are definitely more integrated and follow a completely different approach: a cleaner and more professional integration, which makes the vehicle look like a normal vehicle.
Will the vehicles be unmanned for the whole duration of the trip?
Although the vehicles can run unmanned, they are actually unmanned only during demonstrations. During the trip the vehicles will host people on the back seats as passengers: we are committed to deliver a great show but safety plays an basic role. Therefore people onboard can intervene at any time in case of danger; each intervention is also logged to compute the final percentage of autonomous driving.
Is the solar panel used to recharge the vehicle batteries?
No, the solar panel is used to power the autonomous driving system only. Therefore cameras, lasers, PCs, and actuators are all powered by green energy, making the autonomous driving technology self-sustainable. The ‘autonomous driver’ is therefore seen as a plugin that is completely decoupled form the vehicle system and can be virtually adapted and installed on any vehicle.
How do you recharge the vehicles during the trip? Are there power outlets in the remote areas of the Siberia and China deserts?
Well, there are areas in which it is impossible to find power outlets and therefore recharging the vehicles would be impossible. In these areas we are using power generators. Remember that this is a test: should this be turned into a possible product, power outlets would be disseminated in the area covered by the vehicle.
Which is the autonomy and maximum speed of the vehicles?
The autonomoy of the vehicle and its maximum speed are not affected by our autonomous driver, which is self-sustainable from the point of view of power consumption. Therefore the vehicle perormance are the same as the ones of the original vehicles: about 100 km and 60 km/h.
Assuming it will be a success, what’s left for the future?
A lot! Today we are intervening every day to park the vehicle, recharge it, and perform maintenance. A further challenge would be to let the vehicle run alone for the whole duration of the trip and get it back after 3 months. Something like what happened to our TerraMax vehicle in 2005, when the vehicle was launched in the morning, traveled in the desert alone, and then got back the day after, after spending more than 30 hours of autonomous operations.
So this might be just a first step towards fully autonomous intercontinental driving; more has to be done in the future and if somebody else wants to contact us for a future challenge together we are more than glad to open up our systems and work together.
Are you recording the trip?
Not only we are recording the trip, but we are dumping all data coming from every sensor on the vehicle. This will allow us to run again and again the very same route multimple times, in order to fine tune our algorithms, and also develop new ones.
This will be a unique database, that will include almost every situation related to traffic density, weather, road morphology, road
It is estimated that at the end we will have recorder about 20 Terabytes of unique data!
Which are the most critical issues in this challenge?
Anything is an issue here: from the most trivial aspects such as data recording to the more complex trajectory planning when driving in very congested areas. Even the smallest aspects can become a real issue when the size of the problem gets so large: 13.000 km in 3 months on real roads is a test that has never been tried and documented. Anything can go wrong and therefore it assumes a terrific importance in the research stage. Every little detail has to be planned carefully.