DORA_NAV

DORA_NAV is an open-source autonomous navigation framework built on the DORA dataflow orchestration framework. It supports differential-drive, omnidirectional, and Ackermann-type mobile robot chassis. The framework integrates sensor drivers, mapping, localization, perception, planning, and control into a complete point-to-point navigation pipeline.

The Rerun viewer displays the navigation pipeline in real-time:

Red points — Pre-built PCD point cloud map
Cyan line — Global waypoint path
Blue line — Planned local trajectory (from routing planner)
Green points — Live LiDAR pointcloud (transformed to map frame)
Yellow box — Robot body with heading arrow
Orange trail — Robot trajectory history

Architecture

The system is organized as a DORA dataflow graph where each node runs as an independent process communicating through the DORA runtime.

┌─────────────────────────────────────────────────────────────────────┐
│                        DORA Runtime (v0.3.12)                       │
├─────────────────────────────────────────────────────────────────────┤
│                                                                     │
│  ┌──────────────┐                                                   │
│  │  Sensor Input │   tick (10ms)                                    │
│  │  ────────────│──────────────┐                                   │
│  │  mujoco_sim  │              │                                    │
│  │  (or real    │    ┌─────────▼──────────┐                        │
│  │   LiDAR+IMU) │    │   pointcloud       │                        │
│  │              ├────►   imu_msg           │                        │
│  │              │    │   ground_truth_pose │                        │
│  └──┬───────┬───┘    └──┬──────────┬───────┘                        │
│     │       │           │          │                                │
│     │  SteeringCmd      │     ┌────▼─────────────────┐             │
│     │  TrqBreCmd        │     │  hdl_localization     │             │
│     │   (from control)  │     │  (NDT + UKF)          │             │
│     │                   │     │  ─────────────────    │             │
│     │                   │     │  IN: pointcloud,      │             │
│     │                   │     │      imu_msg          │             │
│     │                   │     │  OUT: cur_pose        │             │
│     │                   │     └────────┬──────────────┘             │
│     │                   │              │                            │
│  ┌──▼───────────┐  ┌───▼──────┐  ┌────▼──────────────────┐        │
│  │  Rerun       │  │ pub_road │  │ road_lane_publisher   │        │
│  │  Visualizer  │  │ (static  │  │ (pose + lane → frenet)│        │
│  │  ──────────  │  │  lanes)  │  │                       │        │
│  │  pointcloud  │  └───┬──────┘  └────────┬──────────────┘        │
│  │  raw_path    │      │                  │                        │
│  │  cur_pose    │      │    road_lane     │  cur_pose_all          │
│  └──────────────┘      │                  │                        │
│                   ┌────▼──────────────────▼───────────────┐        │
│                   │  task_pub_node ──► routing_planning    │        │
│                   │  (road attrs)     (A* + Frenet)       │        │
│                   │                   ────────────────     │        │
│                   │                   OUT: raw_path,       │        │
│                   │                        Request         │        │
│                   └──────┬─────────────────┬──────────────┘        │
│                          │                 │                        │
│                   ┌──────▼──────┐   ┌──────▼──────┐                │
│                   │ lat_control  │   │ lon_control  │               │
│                   │ (Pure       │   │ (PID speed   │               │
│                   │  Pursuit)   │   │  control)    │               │
│                   │ OUT:        │   │ OUT:         │               │
│                   │ SteeringCmd │   │ TrqBreCmd    │               │
│                   └─────────────┘   └──────────────┘               │
│                          │                 │                        │
│                          └────────┬────────┘                        │
│                                   │  (fed back to mujoco_sim       │
│                                   │   or real chassis)              │
│                                   ▼                                 │
│                          ┌────────────────┐                        │
│                          │  Vehicle/Sim   │                        │
│                          └────────────────┘                        │
└─────────────────────────────────────────────────────────────────────┘

Node Summary

Node	Algorithm	Key Libraries	Rate	Description
mujoco_sim	MuJoCo physics + raycasting	MuJoCo 3.5+, OpenMP	1kHz sim, 10Hz LiDAR	Simulated LiDAR, IMU, ground truth pose
hdl_localization	NDT scan matching + UKF	PCL, Eigen3, NDT-OMP	Real-time	Pose estimation against pre-built PCD map
pub_road	Static publisher	DORA C API	5Hz	Publishes pre-loaded road lane geometry
road_lane_publisher	Coordinate transform	Eigen3	Event-driven	Transforms lanes to vehicle-local Frenet frame
task_pub_node	File reader	DORA C API	50Hz	Publishes road attributes (speed limits, stops)
routing_planning	A* + Frenet planning	Eigen3, nlohmann/json	Event-driven	Generates reference trajectory and speed requests
lat_controller	Pure Pursuit	Eigen3	50Hz	Computes steering angle from path
lon_controller	PID	DORA C API	Event-driven	Computes torque/brake from speed request
rerun	Rerun streaming	Rerun SDK, PCL, Eigen3	10Hz	3D visualization of full pipeline state

Shared Data Structures (`include/`)

// SlamPose.h — Localization output
struct Pose2D_h { float x, y, theta; };

// Planning.h — Planning output
struct Request_h {
    uint8_t reques_type;   // FORWARD(0), BACK(1), STOP(2), AEB(3)
    float run_speed;
    float stop_distance;
    float aeb_distance;
};

// Localization.h — Full pose with Frenet coordinates
struct CurPose_h {
    double x, y, theta;   // Cartesian pose
    double s, d;           // Frenet coordinates along lane
};

Directory Structure

dora-nav/
├── build_all.sh                  # Build all nodes
├── dataflow_full_sim.yml         # Full simulation pipeline (run this)
├── run.yml                       # Real hardware pipeline
├── Waypoints.txt                 # Global path waypoints
├── cmake/                        # Shared CMake configuration
│   └── dora_config.cmake
├── include/                      # Shared C/C++ headers
│   ├── SlamPose.h
│   ├── Planning.h
│   ├── Localization.h
│   └── imu_msg.h
├── data/
│   ├── map.pcd                   # Pre-built point cloud map
│   └── path/trajectory.txt       # Recorded trajectory
├── images/
│   └── localization.png          # Rerun visualization screenshot
├── simulation/mujoco_bridge/     # MuJoCo physics simulation
│   ├── src/mujoco_sim_bridge.cpp
│   └── models/robot_warehouse.xml
├── localization/dora-hdl_localization/  # NDT+UKF localization
├── mapping/ndt_mapping/          # Offline NDT map building
├── map/
│   ├── pub_road/                 # Static lane publisher
│   └── road_line_publisher/      # Lane-to-Frenet transformer
├── planning/
│   ├── mission_planning/task_pub/ # Road attribute publisher
│   └── routing_planning/          # A* + Frenet path planner
├── control/vehicle_control/
│   ├── lat_controller/           # Pure Pursuit steering
│   └── lon_controller/           # PID speed control
├── rerun/                        # Rerun 3D visualization
│   ├── src/points_to_rerun.cpp
│   └── data/map.pcd
├── driver/                       # Sensor drivers (LiDAR, IMU, GNSS)
├── peception/                    # Perception (ground filter, clustering, YOLO)
└── vehicle/                      # Chassis drivers (Adora series)

Deployment Guide (Simulation Demo)

Prerequisites

Dependency	Version	Install
Ubuntu	22.04+	-
CMake	3.16+	`sudo apt install cmake`
Clang/GCC	C++17	`sudo apt install clang build-essential`
PCL	Latest	`sudo apt install libpcl-dev`
Eigen3	3.x	`sudo apt install libeigen3-dev`
Boost	Latest	`sudo apt install libboost-all-dev`
yaml-cpp	Latest	`sudo apt install libyaml-cpp-dev`
Python3-dev	3.8+	`sudo apt install python3-dev`
MuJoCo	3.5+	`pip install mujoco`
Rust/Cargo	Latest	`curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs \| sh`

Step 1: Install DORA Framework

# Install DORA CLI
pip install dora-rs-cli

# Clone and build DORA C API library
cd ~/Public
git clone https://github.com/dora-rs/dora.git
cd dora
cargo build -p dora-node-api-c --release
# This produces: ~/Public/dora/target/release/libdora_node_api_c.a
# Headers are in: ~/Public/dora/apis/c/node/

Step 2: Install Rerun Viewer

# Install Rerun SDK (C++ library)
pip install rerun-sdk

# Download and build the C++ SDK
cd /tmp
wget https://github.com/rerun-io/rerun/releases/download/0.29.2/rerun_cpp_sdk.zip
mkdir -p rerun_cpp_sdk && cd rerun_cpp_sdk
unzip ../rerun_cpp_sdk.zip
mkdir -p build && cd build
cmake ..
make -j$(nproc)

Step 3: Clone and Build DORA_NAV

cd ~/Public
git clone https://github.com/RuPingCen/DORA_NAV.git dora-nav
cd dora-nav

Build all pipeline nodes:

./build_all.sh

Or build individual nodes manually:

# Example: build the localization node
cd localization/dora-hdl_localization
mkdir -p build && cd build
cmake .. \
  -DCMAKE_PREFIX_PATH="$HOME/Public/dora-nav/cmake" \
  -DDORA_INCLUDE_DIR="$HOME/Public/dora/apis/c/node" \
  -DDORA_OPERATOR_DIR="$HOME/Public/dora/apis/c/operator" \
  -DDORA_LIB_PATH="$HOME/Public/dora/target/release/libdora_node_api_c.a" \
  -DDORA_NAV_ROOT="$HOME/Public/dora-nav"
make -j$(nproc)

Build the MuJoCo simulation bridge:

cd ~/Public/dora-nav/simulation/mujoco_bridge
mkdir -p build && cd build
cmake .. \
  -DCMAKE_PREFIX_PATH="$HOME/Public/dora-nav/cmake" \
  -DDORA_INCLUDE_DIR="$HOME/Public/dora/apis/c/node" \
  -DDORA_OPERATOR_DIR="$HOME/Public/dora/apis/c/operator" \
  -DDORA_LIB_PATH="$HOME/Public/dora/target/release/libdora_node_api_c.a" \
  -DDORA_NAV_ROOT="$HOME/Public/dora-nav"
make -j$(nproc)

Build the Rerun visualizer:

cd ~/Public/dora-nav/rerun
mkdir -p build && cd build
cmake .. \
  -DCMAKE_PREFIX_PATH="$HOME/Public/dora-nav/cmake" \
  -DDORA_INCLUDE_DIR="$HOME/Public/dora/apis/c/node" \
  -DDORA_OPERATOR_DIR="$HOME/Public/dora/apis/c/operator" \
  -DDORA_LIB_PATH="$HOME/Public/dora/target/release/libdora_node_api_c.a" \
  -DDORA_NAV_ROOT="$HOME/Public/dora-nav"
make -j$(nproc)

Step 4: Verify Builds

Confirm all required binaries exist:

ls -la simulation/mujoco_bridge/build/mujoco_sim_bridge
ls -la localization/dora-hdl_localization/build/hdl_localization
ls -la map/pub_road/build/pubroad
ls -la map/road_line_publisher/build/road_lane_publisher_node
ls -la planning/mission_planning/task_pub/build/task_pub_node
ls -la planning/routing_planning/build/routing_planning_node
ls -la control/vehicle_control/lat_controller/build/lat_controller_node
ls -la control/vehicle_control/lon_controller/build/lon_controller_node
ls -la rerun/build/to_rerun

Step 5: Run the Simulation Demo

Important: Run all dora start commands from the project root directory (dora-nav/) so that relative data file paths resolve correctly.

cd ~/Public/dora-nav

# Start the DORA coordinator and daemon
dora up

# Launch the full simulation pipeline with Rerun visualization
dora start dataflow_full_sim.yml --attach

The Rerun viewer window will open automatically, showing:

The pre-built point cloud map (red)
The global waypoint path (cyan)
Live simulated LiDAR scans (green) transformed by robot pose
The robot body (yellow box) with heading arrow
The robot's trajectory trail (orange)
The planned local path (blue)

Step 6: Stop the Demo

Press Ctrl+C in the terminal, then:

dora destroy

Dataflow Configurations

Dataflow YAML	Description	Sensor Source
`dataflow_full_sim.yml`	Full pipeline with ground truth pose	MuJoCo simulation
`simulation/mujoco_bridge/dataflow_mujoco_sim.yml`	Full pipeline with NDT localization	MuJoCo simulation
`run.yml`	Full pipeline for real robot	Physical LiDAR + IMU

References

DORA Framework — Dataflow-Oriented Robotic Architecture
Rerun — Visualization SDK for robotics
MuJoCo — Physics simulation engine
PCL — Point Cloud Library
HDL Localization — 3D LiDAR localization using NDT
DORA and Rerun installation guide: doc/dora_and_rerun_install.md

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

DORA_NAV

Architecture

Node Summary

Shared Data Structures (`include/`)

Directory Structure

Deployment Guide (Simulation Demo)

Prerequisites

Step 1: Install DORA Framework

Step 2: Install Rerun Viewer

Step 3: Clone and Build DORA_NAV

Step 4: Verify Builds

Step 5: Run the Simulation Demo

Step 6: Stop the Demo

Dataflow Configurations

References

About

Uh oh!

Releases

Packages

Uh oh!

Contributors

Uh oh!

Languages

Name		Name	Last commit message	Last commit date
Latest commit History 19 Commits
apps		apps
chassis		chassis
cmake		cmake
control		control
data		data
doc		doc
driver		driver
images		images
include		include
localization/dora-hdl_localization		localization/dora-hdl_localization
map		map
mapping		mapping
peception		peception
planning		planning
rerun		rerun
simulation/mujoco_bridge		simulation/mujoco_bridge
.gitignore		.gitignore
README.md		README.md
RELEASE.md		RELEASE.md
Waypoints.txt		Waypoints.txt
build_all.sh		build_all.sh
dataflow_full_sim.yml		dataflow_full_sim.yml
load_path.yml		load_path.yml
once_orin.sh		once_orin.sh
road_msg.txt		road_msg.txt
run.yml		run.yml

Folders and files

Latest commit

History

Repository files navigation

DORA_NAV

Architecture

Node Summary

Shared Data Structures (include/)

Directory Structure

Deployment Guide (Simulation Demo)

Prerequisites

Step 1: Install DORA Framework

Step 2: Install Rerun Viewer

Step 3: Clone and Build DORA_NAV

Step 4: Verify Builds

Step 5: Run the Simulation Demo

Step 6: Stop the Demo

Dataflow Configurations

References

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Shared Data Structures (`include/`)

Packages