|
|
b74307c58a
|
feat: Nav2 AMCL integration with VESC odometry + LiDAR (Issue #655)
AMCL-based autonomous navigation on pre-built static maps, wired to
VESC CAN differential-drive odometry (/odom, Issue #646) and RPLiDAR
(/scan) as the primary sensor sources.
New files (saltybot_nav2_slam):
- config/amcl_nav2_params.yaml — complete Nav2 + AMCL parameter file
with inline global/local costmap configs (required by nav2_bringup):
· AMCL: DifferentialMotionModel, 500–3000 particles, z-weights=1.0,
odom_frame=/odom, scan_topic=/scan
· Global costmap: static_layer + obstacle_layer (LiDAR) +
inflation_layer (0.55m radius)
· Local costmap: 4m rolling window, obstacle_layer (LiDAR) +
inflation_layer, global_frame=odom
· DWB controller: 1.0 m/s max, diff-drive constrained (vy=0)
· NavFn A* planner
· Recovery: spin + backup + wait
· Lifecycle managers for localization and navigation
- launch/nav2_amcl_bringup.launch.py — orchestrates:
1. sensors.launch.py (RealSense + RPLIDAR, conditional)
2. odometry_bridge.launch.py (VESC CAN → /odom)
3. nav2_bringup localization_launch.py (map_server + AMCL)
4. nav2_bringup navigation_launch.py (full nav stack)
Exposes: map, use_sim_time, autostart, params_file, include_sensors
- maps/saltybot_map.yaml — placeholder map descriptor (0.05m/cell)
- maps/saltybot_map.pgm — 200×200 P5 PGM, all free space (10m×10m)
- test/test_nav2_amcl.py — 38 unit tests (no ROS2 required):
params structure, z-weight sum, costmap layers, DWB/NavFn validity,
recovery behaviors, PGM format, launch file syntax checks
Updated:
- saltybot_bringup/launch/nav2.launch.py — adds nav_mode argument:
nav_mode:=slam (default, existing RTAB-Map behaviour unchanged)
nav_mode:=amcl (new, delegates to nav2_amcl_bringup.launch.py)
- saltybot_nav2_slam/setup.py — installs new launch, config, maps
- saltybot_nav2_slam/package.xml — adds nav2_amcl, nav2_map_server,
nav2_behaviors, dwb_core, nav2_navfn_planner exec_depends
All 58 tests pass (38 new + 20 from Issue #646).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
|
2026-03-17 11:39:08 -04:00 |
|
|
|
ca95489b1d
|
feat: Nav2 SLAM integration with RPLIDAR + RealSense (Issue #422)
Complete autonomous navigation stack for SaltyBot:
- SLAM Toolbox: online_async 2D LIDAR SLAM from RPLIDAR A1M8
- RealSense D435i: depth → pointcloud → costmap obstacle layer
- Nav2 stack: controllers, planners, behavior server, lifecycle management
- DWB planner: tuned for 20km/h (5.5 m/s) max velocity operation
- VESC odometry bridge: motor telemetry → nav_msgs/Odometry
- Costmap integration: LIDAR + depth for global + local costmaps
- TF tree: complete setup with base_link→laser, camera_link, odom
- Goal interface: /navigate_to_pose action for autonomous goals
Configuration:
- slam_toolbox_params: loop closure, scan matching, fine/coarse search
- nav2_params: AMCL, controllers, planners, behavior trees, lifecycle
- Global costmap: static layer + LIDAR obstacle layer + inflation
- Local costmap: rolling window + LIDAR + RealSense depth + inflation
- DWB planner: 20 vx samples, 40 theta samples, 1.7s horizon
Nodes and launch files:
- vesc_odometry_bridge: integrates motor RPM to wheel odometry
- nav2_slam_bringup: main integrated launch entry point
- depth_to_costmap: RealSense depth processing pipeline
- odometry_bridge: VESC telemetry bridge
Hardware support:
- RPLIDAR A1M8: 5.5 Hz, 12m range, 360° omnidirectional
- RealSense D435i: 15 Hz RGB-D, 200 Hz IMU, depth range 5m
- VESC Flipsky FSESC 4.20: dual motor control via UART
- SaltyBot 2-wheel balancer: 0.35m radius, hoverboard motors
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
|
2026-03-04 23:35:15 -05:00 |
|
