Objective
The goal of PANORAMA is to develop an advanced perception and navigation system for autonomous vehicles, enabling automatic path determination and continuous safe motion in partially structured and partially known environments, such as forests, fields and open mines. The project is mainly concerned with navigation in outdoor environments structured with paths or tracks and some natural or non-natural landmarks.
The goal of PANORAMA is to develop an advanced perception and navigation system for autonomous vehicles, enabling automatic path determination and continuous safe motion in partially structured and partially known environments, such as forests, fields and open mines. The project is mainly concerned with navigation in outdoor environments structured with paths or tracks and some natural or nonnatural landmarks.
Specific objectives are to develop a general purpose perception and navigation system suitable for a large range of outdoor application domains, to test project developments and intermediate results gradually with a fully equipped land vehicle used as a system test bench and to demonstrate capabilities of autonomous motion within real application sites.
The test bed land vehicle now features controlled actuation of steering, throttle and brakes and is able to navigate an outdoor industrial site autonomously by dead reckoning. The following perception tasks have been successfully tested: visual beacons detection for relocalization (colour image processing); obstacle detection up to 30 m ahead of the vehicle (using a laser rangefinder); proximity sensing for close range safety, up to 6 m ahead of the vehicle (using ultrasonic sensors).
The first versions of local path planning, environment modelling and perception planning are now running on a workstation and providing simulation tests. They should be partially integrated onboard the land vehicle test bed in the near future.
In addition, the project has issued an original computing hardware architecture that allies the advantages of well known VME-based systems (for interfaces with sensors and actuators) and the high computing power potentialities of transputer based architectures (parallelization of high level processing for modelling, planning and decision processes), enabling an easy upgrading of system performance and a large degree of development flexibility.
Specific objectives are to:
- develop a general-purpose perception and navigation system suitable for a large range of outdoor application domains
- gradually test project developments and intermediate results with a fully equipped land vehicle used as a system test-bench
- demonstrate capabilities of autonomous motion within real application sites
The navigation technique relies on a combination of proprioceptive and exteroceptive data. The approach incorporates both dead reckoning and external cues, such as global positioning systems and visual markers (eg road edges and landmarks). An inertial navigation system is used in combination with a database modelling the road network and landmarks.
Fields of science
- engineering and technologymechanical engineeringvehicle engineeringautomotive engineeringautonomous vehicles
- engineering and technologymaterials engineeringcolors
- engineering and technologyelectrical engineering, electronic engineering, information engineeringelectronic engineeringsensors
- social sciencessocial geographytransportnavigation systemsinertial navigation system
- natural sciencesphysical sciencesopticslaser physics
Topic(s)
Data not availableCall for proposal
Data not availableFunding Scheme
Data not availableCoordinator
75783 Paris
France