Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
IKBT: Solving Symbolic Inverse Kinematics with Behavior Tree
Authors: Dianmu Zhang, Blake Hannaford
JAIR 2019 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We tested IKBT on many sets of DH parameters, representing serial arm robot designs (including commercial robots, and solved design examples from student homework), the successful solving rate is listed in Table 1. As the DOF number increases, the problem becomes more complex and the success rate decreases. In general it solves most of the robots, up to 6 DOF. Note that IKBT can solve robots regardless of their configurations, e.g. IKBT does not require robots having three intersecting axes. [...] IKBT successfully solves complicated robots, such as the 6-DOF commercial robot manipulator PUMA 560 and successfully solved 18 out of 19 test robots (95% success rate) (see Results section). |
| Researcher Affiliation | Academia | Dianmu Zhang EMAIL Blake Hannaford EMAIL Electrical and Computer Engineering University of Washington Seattle, WA, 98195 USA |
| Pseudocode | No | The paper describes the system's workflow, architecture, transformations, and solvers in detailed prose and includes a structural diagram (Figure 2: IKBT Structure). However, it does not present any formal pseudocode or algorithm blocks with structured steps in a code-like format. |
| Open Source Code | Yes | Source code can be found at: https://github.com/uw-biorobotics/IKBT. The DH parameters of all these robots are stored as part of the source code (in ik_robots.py), for purpose of testing and reproducing the results. Instructions are on the Git Hub page. [...] Full source code and test case examples are available at Github: https://github.com/uw-biorobotics/IKBT. |
| Open Datasets | Yes | Source code can be found at: https://github.com/uw-biorobotics/IKBT. The DH parameters of all these robots are stored as part of the source code (in ik_robots.py), for purpose of testing and reproducing the results. Instructions are on the Git Hub page. |
| Dataset Splits | No | The paper mentions testing IKBT on 'many sets of DH parameters' and '18 out of 19 test robots' but does not provide specific details on dataset splits (e.g., percentages, sample counts for training/validation/testing, or explicit splitting methodology). |
| Hardware Specification | No | On average, IKBT generates symbolic solutions and source code in a few minutes on a normal PC. |
| Software Dependencies | No | IKBT requires few dependencies outside of the standard Python distribution (mainly the symbolic manipulation package sympy and the unit testing framework unittest). |
| Experiment Setup | Yes | To verify the solution, we followed the process stated in section 5.1. The starting pose used is: θ1 = 30 , θ2 = 50 , θ3 = 40 , θ4 = 45 , θ5 = 120 , θ6 = 60 as well as the parameters: a2 = 5, a3 = 1, d3 = 2, d4 = 4 |