knowrob_ros

knowrob_ros is a ROS 1 wrapper for KnowRob, providing ROS interfaces to integrate KnowRob's knowledge reasoning capabilities with robotic systems.

Key Features

• ROS Actions
  • AskOne, AskAll, AskIncremental, and Tell
• Python API
  • knowrob_ros_lib.py for simple query/assertion via ros in python
• Docker Support
  • Ready-to-use container for fast setup

Installation (Local ROS Workspace)

Prerequisites

• Ubuntu 20.04
• ROS Noetic

For the rest, see the dependencies of KnowRob.

1. Clone the Repositories

mkdir -p ~/catkin_ws/src
cd ~/catkin_ws/src

# Clone KnowRob
git clone https://github.com/knowrob/knowrob.git

# Clone knowrob_ros (this repo)
git clone https://github.com/knowrob/knowrob_ros.git

2. Build the workspace

cd ~/catkin_ws
catkin build

---

Installation with Docker

If you'd like to avoid setting up everything locally, you can use the provided Dockerfile to run KnowRob and knowrob_ros in an isolated container.

1. Build the Docker Image

From the root of the repository:

docker build -t knowrob/ros1 .

2. Run the Container

docker run -it --entrypoint bash knowrob/ros1

3. Launch KnowRob in Docker

Once inside the container:

source /catkin_ws/devel/setup.bash
roslaunch knowrob_ros knowrob.launch

Usage

ROS Interfaces

The following ROS interfaces are available:

Endpoint	Type	Description
/knowrob/askone	AskOneAction	Single-result query
/knowrob/askall	AskAllAction	Multi-result query
/knowrob/askincremental	AskIncrementalAction	Start incremental query
/knowrob/askincremental_next_solution	AskIncrementalNextSolutionAction	Fetch next solution
/knowrob/askincremental_finish (service)	AskIncrementalFinish.srv	Finish incremental query
/knowrob/tell	TellAction	Assert triples into KB

All endpoints need to be called with a ModalFrame message. We document the ModalFrame message in detail below.

Using the Interface from Python

Use the Python client library to integrate KnowRob queries and assertions directly into your code. Here’s a minimal example:

from knowrob_ros.knowrob_ros_lib import KnowRobRosLib, get_default_modalframe

# Initialize ROS node and client
client = KnowRobRosLib()
client.init_node("my_python_client")

# Single-result query
modal = get_default_modalframe()
response = client.ask_one("lpn:jealous(lpn:vincent, X)", modal)
print("AskOne status:", response.status)
print("Binding:", response.answer)

# Shutdown when done
client.shutdown_node()

For detailed usage, full method reference, and advanced examples, see the Python API Guide.

Command Line

To run a query insterface on the terminal (for docker you can connect to the same container in which you launched knowrob), you should first list to the corresponding topic, on which the results are published:

rostopic echo /knowrob/askone/result 

Then run a query on another terminal:

rostopic pub /knowrob/askone/goal knowrob_ros/AskOneActionGoal "header:
  seq: 0
  stamp: {secs: 0, nsecs: 0}
  frame_id: ''
goal_id:
  stamp: {secs: 0, nsecs: 0}
  id: ''
goal:
  query:
    lang: ''
    queryString: 'lpn:jealous(lpn:vincent, X)'
    frame:
      epistemicOperator: 0                                                                                                          
      aboutAgentIRI: ''
      aboutSimulationIRI: ''                                                             
      temporalOperator: 0
      minPastTimestamp: 0.0
      maxPastTimestamp: 0.0
      confidence: 0.0"

You should now see an answer on the first terminal.

Modal Frames

The ModalFrame message controls who “knows” what, and when it’s considered true. Supply one on every call to ask_one, ask_all, ask_incremental, or tell.

Field	Type	Default	Description
epistemicOperator	uint8	KNOWLEDGE (0)	KNOWLEDGE: “it is certain that …” BELIEF (1): “it could be that …”
aboutAgentIRI	string	""	IRI of another agent whose knowledge/belief you’re querying. Mutually exclusive with aboutSimulationIRI.
aboutSimulationIRI	string	""	IRI of a mental simulation knowledge graph to query. Mutually exclusive with aboutAgentIRI.
temporalOperator	uint8	CURRENTLY (0)	CURRENTLY: “it is now the case that …” ALL_PAST (1): “it always was the case that …” SOME_PAST (2): “it was the case that …”
minPastTimestamp	float	UNSPECIFIED_TIMESTAMP (-1.0)	Lower time bound (seconds since epoch). When >0, only facts true after this timestamp are considered.
maxPastTimestamp	float	UNSPECIFIED_TIMESTAMP (-1.0)	Upper time bound (seconds since epoch). When >0, only facts true before this timestamp are considered.
confidence	float	0.0	Minimum confidence threshold (0.0–1.0) for returned statements.

Repository Layout

.
├── action/            ← ROS action definitions
├── include/           ← C++ interface headers
├── launch/            ← Launch files
├── msg/               ← Custom ROS messages
├── scripts/           ← Python test scripts
├── src/               ← Core Prolog, C++ and Python code
│   └── knowrob_ros_lib/ ← Python client library
├── srv/               ← ROS services
├── test/              ← rostest definitions
├── Dockerfile         ← Container setup
├── build-docker.sh    ← Helper for Docker builds
├── CMakeLists.txt
├── package.xml
└── README.md

Contributing

1. Fork the repo and create a branch
2. Write or update tests in scripts/ or test/
3. Follow ROS and Python style guidelines
4. Submit a pull request

Support

• Issues & feature requests: GitHub Issues

(Inactive issues will be closed after 2 weeks of inactivity. Please open a new issue if you need further assistance.)