Vision-guided, autonomous vans travel from Italy to China
After 13,000 km and three months through Europe and Asia, a convoy of unmanned, solar-energy-powered vans crossed the Chinese border. The expedition started on 20 July in Italy and the goal is to reach the World Expo in Shanghai on 28 October, where a demonstration will take place at the European Union/Belgian Pavilion.
The Intercontinental Autonomous Challenge is part of a research project on autonomous driving made possible by significant funding by the European Research Council (ERC). The van is equipped with front and rear color stereo vision systems, multiple laser scanners, and a panoramic vision system that stitches together images from three synchronized cameras that can track the leader vehicle in the convoy.
"We have now covered more than two thirds of the route. When it comes to the technology of the vehicles it went smoothly all in all," said lead researcher Alberto Broggi from the Vislab at Parma University, Italy. He added: “We are happy about the test and for the great amount of data we are acquiring for further processing at the end of the trip.
The convoy, which includes back-up vehicles, has travelled about 6500 km in 80 days and now passed the border between Kazakhstan and China. The remaining part of the trip will go through varied Chinese landscape, including the Gobi desert.
Along the varied route to Shanghai, the vans are faced with a range of different environmental and traffic conditions, including extreme ones. The trip is part of an experiment testing sensory technology and will show that it is possible--although only in a prototype version--to move goods between two continents with non-polluting vehicles and with virtually no human intervention.
Vehicle behavior
The vehicles have to face a plethora of extremely different and extreme conditions: road, weather, infrastructures, temperature, traffic, and even possibly unlawful behavior of other road participants. Plus, the route is unknown since no maps of a large percentage of the trip are available.
To face these issues, VisLab put extra care in the design of the vehicles and in the definition of their behavior: two autonomous vehicles are moving at the same time during the trip. Although the two vehicles are exactly identical (same sensor suite and identical control software) they have different goals.
The first vehicle drives autonomously for most of the trip; it conducts experimental tests on sensing, decision, and control subsystems, and collects data throughout the whole trip. Although limited, human interventions are needed to define the route and intervene in critical situations.
The second vehicle automatically follows the route defined by the preceding vehicle, requiring no human intervention (100% autonomous). It will be regarded as a readily exploitable vehicle, able to move on loosely predefined routes. At the end of the trip, its technology -thoroughly assessed and refined during this extreme trip- will be mature enough to be used on a set of vehicles moving downtown.
Leader-Follower
The two vehicles in motion follow an enhanced leader-follower approach:
• in case the leader is visible, the follower locates the leader and follows it; local sensing is used to refine its position on the road, avoid obstacles, and determine speed.
• in case the leader is not visible, the follower uses GPS coordinates sent by the leader to determine a rough route; local sensing is again used to refine its position on the road, avoid obstacles, determine speed, and follow the road/path.
To read more, see videos, and follow the blog, visit the Vislab website.
Posted by Conard Holton
Vision Systems Design