SENSE. PLAN. ACT.
Mobile Robot Architecture
From sensor observations to supervised motion. Follow how perception, state estimation, planning and control connect the robot to its environment.
Mobile robot architectureC / C++
MissionDestination / task
Prior knowledgeMap · traffic rules · sign database
↓ Supervised actuator commands↑ Sensor observations + motion feedback
Physical executionSEPARATE DOMAIN
Real robot or simulated plant
↔Drive · steering · brakes · chassis
Vehicle dynamics can be implemented in Simulink.Environment
Terrain · weather · other participants
Illustrative functional architecture. Localization estimates pose against a known map; SLAM estimates pose and map together. The physical robot or simulated plant is separate from the autonomy software.