Planning with Task-Oriented Knowledge Acquisition for a Service Robot
Authors: Kai Chen, Fangkai Yang, Xiaoping Chen
IJCAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate the approach on service robot Ke Jia that serves drink to guests, a testing benchmark for general-purpose service robot proposed by Robo Cup@Home competition.The experiment results are shown in Section 7.Following the scoring policy of Robo Cup@Home competition, we evaluate our framework by conducting a total of over 20 hours running the robot and finished about 157 orders in 45 trials. |
| Researcher Affiliation | Collaboration | 1,3School of Computer Science and Technology, University of Science and Technology of China 96 Jinzhai Rd, Hefei, Anhui, 200027, China 1chk0105@mail.ustc.edu.cn, 3xpchen@ustc.edu.cn 2 Katy Drilling Software Center, Schlumberger Software Technology, Schlumberger Ltd. 23500 Colonial Parkway, Katy, TX, 77493, USA fyang10@slb.com |
| Pseudocode | Yes | The main control loop for plan generation and execution is presented in Algorithm 1.Continuous sensing (Algorithm 2)...Execution monitor dispatches action commands to corresponding controllers in robot interface for execution. Controllers operate hardware and, like in continuous observation, ground and return symbolic results (Algorithm 3). |
| Open Source Code | No | Our demo video (https://youtu.be/zoKNFozlFPk) shows 7 consecutive tasks in one trial how robot responds to so many challenges. (This is a video link, not a code link.) |
| Open Datasets | No | No specific dataset is mentioned with access information. The paper describes evaluation on a service robot in the context of the Robo Cup@Home competition, running custom scenarios. |
| Dataset Splits | No | The paper does not provide specific dataset split information (e.g., train/validation/test percentages or counts). Experiments were conducted on a physical robot Ke Jia in real-world scenarios. |
| Hardware Specification | Yes | Ke Jia (Figure 3(a)) is equipped with a wheeled mobile base, a single 5-Degree-of-Freedom arm, a microphone, 2D laser range finder, Microsoft Kinect and a high-resolution camera. |
| Software Dependencies | No | No specific version numbers for software dependencies are provided. The paper mentions 'Our system is implemented as a node in ROS (Robot Operating System) network.' and 'In experiment we use CLASP as our answer set solver'. |
| Experiment Setup | Yes | The initial fluent set S for planning is generated as follows (line 1): (i) fluents that belong to definite world state are initialized based on robot s own sensor inputs; (ii) fluents that belong to belief state are initialized as negated literals, denoting that the robot does not know anything about them. Possible world state is not specified for initial state. |