{"id":19134202,"url":"https://github.com/um-arm-lab/traversability-based-contact-space-planner","last_synced_at":"2026-06-15T06:33:42.390Z","repository":{"id":77983860,"uuid":"237724878","full_name":"UM-ARM-Lab/Traversability-Based-Contact-Space-Planner","owner":"UM-ARM-Lab","description":"Traversability-based contact planner code release","archived":false,"fork":false,"pushed_at":"2020-02-12T15:43:11.000Z","size":79155,"stargazers_count":4,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"master","last_synced_at":"2025-11-13T06:03:57.609Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"C++","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/UM-ARM-Lab.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2020-02-02T05:43:36.000Z","updated_at":"2025-10-29T19:22:28.000Z","dependencies_parsed_at":"2023-05-14T09:15:35.681Z","dependency_job_id":null,"html_url":"https://github.com/UM-ARM-Lab/Traversability-Based-Contact-Space-Planner","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/UM-ARM-Lab/Traversability-Based-Contact-Space-Planner","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/UM-ARM-Lab%2FTraversability-Based-Contact-Space-Planner","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/UM-ARM-Lab%2FTraversability-Based-Contact-Space-Planner/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/UM-ARM-Lab%2FTraversability-Based-Contact-Space-Planner/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/UM-ARM-Lab%2FTraversability-Based-Contact-Space-Planner/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/UM-ARM-Lab","download_url":"https://codeload.github.com/UM-ARM-Lab/Traversability-Based-Contact-Space-Planner/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/UM-ARM-Lab%2FTraversability-Based-Contact-Space-Planner/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":34351448,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-06-15T02:00:07.085Z","response_time":63,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2024-11-09T06:25:51.837Z","updated_at":"2026-06-15T06:33:42.371Z","avatar_url":"https://github.com/UM-ARM-Lab.png","language":"C++","funding_links":[],"categories":[],"sub_categories":[],"readme":"Traversability-based Contact Space Planner\n=============\n\nTraversability-based contact space planner code release. Given environment specified as a set of polygonal surface and a robot model, the code generate contact sequence for a humanoid robot. We provide an example with the [Escher humanoid robot](https://icat.vt.edu/projects/2015-2016/major/escher-humanoid-robot.html) model. The code is written in Python 2.7 and tested in Ubuntu 14.04 with ROS Indigo.\n\nSetup\n------------\n* Install [OpenRAVE](https://github.com/rdiankov/openrave), [Installation Guide](https://scaron.info/teaching/installing-openrave-on-ubuntu-14.04.html) (Tested Commit: 7c5f5e27eec2b2ef10aa63fbc519a998c276f908)\n* Install [ROS](http://wiki.ros.org/indigo/Installation/Ubuntu)\n* Install `srdfdom` and `tinyxml2`: `sudo apt-get install ros-indigo-srdfdom libtinyxml2-dev`\n* Install `Eigen` and `NEWMAT`: `sudo apt-get install libeigen3-dev libnewmat10-dev`\n* Install [cddlib](https://inf.ethz.ch/personal/fukudak/cdd_home/) (Tested Version: cddlib-094h):\n\n  Get cddlib from ftp:\n  ```\n  wget ftp://ftp.math.ethz.ch/users/fukudak/cdd/cddlib-094h.tar.gz\n  tar -xvf cddlib-094h.tar.gz\n  cd cddlib-094h\n  ```\n  or from github:\n  ```\n  git clone https://github.com/cddlib/cddlib.git\n  cd cddlib\n  git checkout 0.94h\n  ```\n  Install cddlib and dependencies:\n  ```\n  sudo apt-get install libgmp3-dev\n  ./configure\n  make\n  sudo make install\n  cd /usr/local/include\n  sudo mkdir cdd\n  sudo mv cdd_f.h cddmp_f.h cddtypes_f.h cdd.h cddmp.h cddtypes.h setoper.h cdd\n  ```\n\n* Create a catkin workspace and clone the repo under `/path/to/catkin/workspace/src`, run `catkin_make` in `/path/to/catkin/workspace` to build the code.\n* Before running the code, `source /path/to/catkin/workspace/devel/setup.bash`.\n\nUsage\n-----\n\nThe contact planner is initiated with the script `humanoid_motion_planner.py` with the following options:\n\n* `contact_sequence_generation_method`: Decide how contact sequence is generated for each guiding path segment.\u003cbr/\u003e\n  `all_planning`: Use graph search planning in every segment.\u003cbr/\u003e\n  `all_retrival`: Retrive and adapt previously generated motion plan in evey segment.\u003cbr/\u003e\n  `hybrid`(Default): Use planning in segment with high traversability, and use retrival in segment with low traversability.\n\n* `path_segmentation_type`: Decide how guiding path is segmented.\u003cbr/\u003e\n  `no_segmentation`: The planner will use every motion mode along the guiding path.\u003cbr/\u003e\n  `motion_mode_segmentation`: The planner will segment the guiding path at where motion mode changes.\u003cbr/\u003e\n  `motion_mode_and_traversability_segmentation`(Default): The planner will segment the guiding path at where motion mode changes, and then further decompose each segment based on the traversability.\n\n* `traversability_threshold_type`: The type of threshold that the planner use to determine the contact sequence generation method.\u003cbr/\u003e\n  `mean`(Default): The planner uses the mean of the traversability of all torso transition in a segment to determine using planning or retrival method to generate contact sequence in the segment.\u003cbr/\u003e\n  `max`: The planner uses the max of the traversability of all torso transition in a segment to determine using planning or retrival method to generate contact sequence in the segment.\n\n* `surface_source`: The source of the environment in planning. The repo provides 3 examples each for two-corridor and two-stair environment. The user can create new environment by adding new options in `update_environment` function in `environment_handler.py`.\u003cbr/\u003e\n  `two_corridor_environment`(Default): Randomly generate a two-corridor environment.\u003cbr/\u003e\n  `two_stair_environment`: Randomly generate a two-stair environment.\u003cbr/\u003e\n  `load_from_data`: Load environment object file stored using `pickle` from path specified by `environment_path` parameter.\n\n* `environment_path`: The folder which contains the stored environment object file.\n* `start_env_id` and `end_env_id`: The first and last environment object file id loaded in the process.\n\n\nExample Usage:\n```\npython humanoid_motion_planner.py contact_sequence_generation_method hybrid traversability_threshold_type mean path_segmentation_type motion_mode_and_traversability_segmentation environment_path environment_two_corridor surface_source load_from_data start_env_id 0 end_env_id 0\n```\n\nUsing A Different Robot Model\n-----\n\nTo swap a robot model, check the the comments in `load_escher.py`, and define all corresponding variables for the new robot model to create a robot loading function. Then swap all `load_escher` functions with the newly defined robot loading function.\n\nTraining Traversability Regressor\n-----\n\nTo train traversability regressor, first generate the footstep_window, which precomputes all the footstep location given a torso transition. Then collect training data for each separate regressor mode (legs_only, legs_and_one_hand, all_manipulators). Finally, train the model with the training data. In summary, run the scripts with the following steps:\n1. `python transition_footstep_window_generator.py` (Generate footstep_window)\n2. `python traversability_training_data_collector.py batch_id [training data batch id] mode [legs_only,legs_and_one_hand,all_manipulators] sample_env_num [number of sampling environment] surface_source [environment name]` (Collect training data for each mode.)\n3. `python traversability_regressor_training.py mode [legs_only,legs_and_one_hand,all_manipulators]` (Train traversability regressor for each mode.)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fum-arm-lab%2Ftraversability-based-contact-space-planner","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fum-arm-lab%2Ftraversability-based-contact-space-planner","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fum-arm-lab%2Ftraversability-based-contact-space-planner/lists"}