Index

Detection Area

Role

If pointcloud or predicted objects are detected in a detection area defined on a map, the stop planning will be executed at the predetermined point.

Activation Timing

This module is activated when there is a detection area on the target lane.

Module Parameters

Parameter	Type	Description
use_dead_line	bool	[-] weather to use dead line or not
state_clear_time	double	[s] when the vehicle is stopping for certain time without incoming obstacle, move to STOPPED state
stop_margin	double	[m] a margin that the vehicle tries to stop before stop_line
dead_line_margin	double	[m] ignore threshold that vehicle behind is collide with ego vehicle or not
unstoppable_policy	string	[-] policy for handling unstoppable situations: "go" (pass through), "force_stop" (emergency stop), or "stop_after_stopline" (stop after the stop line)
max_deceleration	double	[m/s^2] maximum deceleration used to calculate required braking distance for unstoppable situation handling
delay_response_time	double	[s] delay response time used to calculate required braking distance for unstoppable situation handling
hold_stop_margin_distance	double	[m] parameter for restart prevention (See Algorithm section)
distance_to_judge_over_stop_line	double	[m] parameter for judging that the stop line has been crossed
suppress_pass_judge_when_stopping	bool	[m] parameter for suppressing pass judge when stopping
enable_detected_obstacle_logging	bool	[-] enable/disable logging of detected obstacle positions, time elapsed since last detection, and ego vehicle position when ego-vehicle is in STOP state
target_filtering.pointcloud	bool	[-] whether to stop for pointcloud detection
target_filtering.unknown	bool	[-] whether to stop for UNKNOWN objects area
target_filtering.car	bool	[-] whether to stop for CAR objects area
target_filtering.truck	bool	[-] whether to stop for TRUCK objects area
target_filtering.bus	bool	[-] whether to stop for BUS objects area
target_filtering.trailer	bool	[-] whether to stop for TRAILER objects area
target_filtering.motorcycle	bool	[-] whether to stop for MOTORCYCLE objects area
target_filtering.bicycle	bool	[-] whether to stop for BICYCLE objects area
target_filtering.pedestrian	bool	[-] whether to stop for PEDESTRIAN objects area
target_filtering.animal	bool	[-] whether to stop for ANIMAL objects area
target_filtering.hazard	bool	[-] whether to stop for HAZARD objects area
target_filtering.over_drivable	bool	[-] whether to stop for OVER_DRIVABLE objects area
target_filtering.under_drivable	bool	[-] whether to stop for UNDER_DRIVABLE objects area

Inner-workings / Algorithm

1. Gets a detection area and stop line from map information and confirms if there are obstacles in the detection area
2. Inserts stop point l[m] in front of the stop line
3. Calculates required braking distance based on current velocity, max_deceleration, and delay_response_time
4. If the vehicle cannot stop before the stop line, applies the configured unstoppable_policy
5. Sets velocity as zero at the determined stop point

Detection Logic

The module uses two detection sources:

• Pointcloud detection: Detects any obstacles in the detection area using 3D point cloud data (if target_filtering.pointcloud is enabled)
• Predicted objects detection: Detects classified objects (vehicles, pedestrians, etc.) in the detection area based on perception module outputs

The module stops the vehicle if either detection source finds an obstacle. For performance optimization, if pointcloud detection finds an obstacle, predicted objects detection is skipped (short-circuit evaluation).

Flowchart

Restart prevention

If it needs X meters (e.g. 0.5 meters) to stop once the vehicle starts moving due to the poor vehicle control performance, the vehicle goes over the stopping position that should be strictly observed when the vehicle starts to moving in order to approach the near stop point (e.g. 0.3 meters away).

This module has parameter hold_stop_margin_distance in order to prevent from these redundant restart. If the vehicle is stopped within hold_stop_margin_distance meters from stop point of the module (_front_to_stop_line < hold_stop_margin_distance), the module judges that the vehicle has already stopped for the module's stop point and plans to keep stopping current position even if the vehicle is stopped due to other factors.

parameters

outside the hold_stop_margin_distance

inside the hold_stop_margin_distance

Unstoppable situation handling

When the ego vehicle cannot stop before the stop line with the given max_deceleration and delay_response_time, the module applies the unstoppable_policy:

• "go" policy: The vehicle is allowed to pass through without stopping. A warning is logged.
• "force_stop" policy: The vehicle performs a stop at the original stop line, even if it cannot stop comfortably. A warning is logged.
• "stop_after_stopline" policy: The stop point is shifted forward beyond the stop line to ensure the vehicle can stop safely within the physical limits.

The required braking distance is calculated as: \(d*{req} = v \cdot t*{delay} + v^2/(2a*{max})\), where \(v\) is current velocity, \(t*{delay}\) is delay_response_time, and \(a\_{max}\) is max_deceleration.

For the "stop*after_stopline" policy, if \(d*{req}\) exceeds the remaining distance to the stop line \(d*{stop}\), the stop point is shifted forward by \(d*{req} - d\_{stop}\). This adjustment is applied only once, when the module transitions from GO to STOP state.