slam

slam.py

Overview

slam.py is a minimal SLAM façade that publishes a blank occupancy map and identity odometry so that navigation stacks (e.g. move_base, rviz) can start before a real SLAM backend is available. It defines exactly which message types and frames future SLAM solutions must honour.

Interfaces (strongly-typed, stateless)

Direction / Type	Topic	Semantics
Required (sub)	/{robot}/fused_scan	sensor_msgs/LaserScan – pre-filtered scan from sensor_fusion
Required (sub)	/{robot}/odom_proc	nav_msgs/Odometry – upstream pose estimate (ignored by stub)
Provided (pub, latched)	/{robot}/map	nav_msgs/OccupancyGrid – blank grid, 0.05 m cells, all free/unknown
Provided (pub, latched)	/{robot}/odom_slam	nav_msgs/Odometry – identity pose, zero twist, frame_id odom

Coordinate frames

map ←— identity —→ odom ←— real TF —→ base_link No TF broadcaster here; downstream must supply static or real transforms.

Contract

Pre-conditions

• Nodes publish /fused_scan at ≥ 1 Hz.

• /odom_proc provides any Odometry messages (not used by stub).

Post-conditions

• On the first incoming scan or odom message each publisher latches and makes the blank map and identity odom available.

• No further updates unless subscribers re-connect.

Implementation notes

• Map: 10 × 10 cells, 0.05 m resolution, origin (–2.5, –2.5) so (0,0) is map centre.

• Odometry: pose = (0,0,0), orientation = unit quaternion, twist = zero.

• Replace update_map and update_pose stubs to integrate real SLAM algorithms.

Classes

SLAM()

Stub container wiring subscribers and latched publishers.