Planning and Control

The engineering layer, and job title, covering both deciding how to move and making the actuators follow that decision.

‘Planning and Control,’ abbreviated PnC, is common shorthand in Chinese industry for the layer that sits between perception and the actuators. The planning half turns localization, perception results, and the task goal into a path and a timed trajectory; the control half computes motor or actuator commands from the current state so the robot tracks that trajectory. The term is especially common in the self-driving industry: Baidu Apollo splits its system into a planning module (which outputs a trajectory) and a control module (which computes steering, throttle, and braking commands using LQR, PID, or MPC). At robotics companies, a PnC team is typically responsible for motion planning, trajectory optimization, MPC, and whole-body control — roughly the ‘cerebellum’ half of a brain-cerebellum split. End-to-end models now hand part of the planning job to a neural network, but low-level control and safety constraints are still mostly handled by PnC.

ExampleIn Apollo, the planning module outputs a driving trajectory with speed and acceleration profiles; the control module uses LQR for lateral steering and PID for longitudinal throttle and braking to keep the car on that trajectory.

Also called

PnC, Planning & Control

Related

Motion Planning · Trajectory Planning · Model Predictive Control · Whole-Body Control · Brain–Cerebellum Architecture · Autonomous Driving

Sources

Apollo Planning 模块说明（README_cn） (Chinese)
Apollo Control 模块说明（README_cn） (Chinese)