Integrated Motion Planning for Assembly Task with Part Manipulation Using Re-Grasping

Round 1

Reviewer 1 Report

Thank you for the opportunity to review this manuscript.
Authors present a possible solution for planning motion during assembly tasks. The designed planner is based on RRT (Rapidly Exploring Random Tree), and authors consider the re-grasping opportunity. Given multiple grasps poses possibilities for the part of assembling, the designed planner chooses candidate grasp poses considering the environment. This is a common problem, and the adopted solution is readily implementable on the existing manipulator systems without any modifications, so this topic has value to the community.

The authors present enough review of the bibliography, showing ways to introduce and study the problem. The starting citations in the paper could be improved, adding some other cases could benefit of the present study, i.e. adding one of the follows:

Conte, G., Scaradozzi, D., Casalino, G., Simetti, E., & Sperindè, A. (2016, June). A Robotic Platform for Underwater Assisted Manipulation. In The 26th International Ocean and Polar Engineering Conference. International Society of Offshore and Polar Engineers.

Barbieri, L., Bruno, F., Gallo, A., Muzzupappa, M., Russo, M.L., “Design, prototyping and testing of a modular small-sized underwater robotic arm controlled through a Master-Slave approach” (2018) Ocean Engineering, 158, pp. 253-262.

The reviewer kindly suggests adding a table with pros and specifications of PRM, RRT, EST, SBL and other techniques authors think could be useful compare for a better reading of the introduced problem.
The results discussion must be improved with more cases and comparing the times with well-known cases from literature (please improve the Table 1).

Commenting conclusion and future work the authors kindly suggest to consider the case when the distance of the obstacle can be measured. See, i.e. the following:

Novak, J. L., & Feddema, J. T. (1991). A capacitance-based proximity sensor for whole arm obstacle avoidance (No. SAND-91-2119C; CONF-920540-14). Sandia National Labs., Albuquerque, NM (United States).

De Leo, A., Scaradozzi, D., Genovesi, R., Cerri, G., Conte, G., Perdon, A. M., & Omerdic, E. (2019, June). Preliminary Study of a Novel Magnetic Sensor for Safety in Industrial Robotics. In 2019 IEEE International Symposium on Robotic and Sensors Environments (ROSE) (pp. 1-6). IEEE.

Conte, G., Scaradozzi, D., & Rosettani, M. (2011). E-field Sensors and Sensor-based Control Strategies for M/M Safe Cooperation. IFAC Proceedings Volumes, 44(1), 8070-8075.

Mayton, B., LeGrand, L., & Smith, J. R. (2010, May). An electric field pretouch system for grasping and co-manipulation. In 2010 IEEE International Conference on Robotics and Automation (pp. 831-838). IEEE.

In summary, this paper ends up showing that build a modular and lightweight strategy to plan a path is feasible. Therefore, my recommendation is to accept the manuscript with minor revision.

Author Response

Please find the attached file below.

Author Response File: Author Response.pdf

Reviewer 2 Report

This work focusses on an integrated grasp and motion planning algorithm for robotic assembly. The work assumes that a set of feasible grasps is already given and, takes into account the geometry of parts and their initial, and final poses to choose feasible grasps. The proposed algorithm is validated to assemble the SOMA puzzle with a 6DOF UR5e. The authors claim that by using an Orientation Graph Search based re-grasping approach in task space, the proposed method is computationally less costly than the configuration-space search-based approaches.

The novelty of this paper is unclear. Integrated solutions for grasp and motion planning have been there for multiple decennia and are used in commercial contexts. Today, robots can grasp and manipulate much more complex objects in an industrial setting. The literature study in this paper is limited and authors failed to integrate and compare with similar work: see [4-5] for a survey and [1-3] for similar approaches. According to [6] current challenges for robotic assembly are on fine assembly as well on data-driven task learning and perception.

In the results section, computation time in seconds of planning for 3 different assembly formations is given. How does this compare with state-of-the-art commercial and experimental software packages with similar functionality? What was the computing device?

Why are snap-shots used to show the progress of the robot? A video would be much more informative.

References

[1] Wan, Weiwei, Kensuke Harada, and Kazuyuki Nagata. "Assembly sequence planning for motion planning." Assembly Automation 38.2 (2018): 195-206.

[2] Bagnell, J. Andrew, et al. "An integrated system for autonomous robotics manipulation." 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems. IEEE, 2012.

[3] Wan, Weiwei, and Kensuke Harada. "Integrated assembly and motion planning using regrasp graphs." Robotics and biomimetics 3.1 (2016): 18.

[4] Ghandi, Somaye, and Ellips Masehian. "Review and taxonomies of assembly and disassembly path planning problems and approaches." Computer-Aided Design 67 (2015): 58-86. 

[5] Jiménez, Pablo. "Survey on assembly sequencing: a combinatorial and geometrical perspective." Journal of Intelligent Manufacturing 24.2 (2013): 235-250.

[6] Zhu, Zuyuan, and Huosheng Hu. "Robot learning from demonstration in robotic assembly: A survey." Robotics 7.2 (2018): 17.

Author Response

Please find the attached file below.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The authors improved the literature study and motivation of the manuscript. However, the authors fail to respond to:

How does this compare with state-of-the-art commercial and experimental software packages with similar functionality? 

Inclusion of a video to support the snapshots and the results

 

Author Response

Respected Reviewer,

We are very much thankful to you for your deep and thorough review. We found your comments very helpful for our research work. We have revised our manuscript according to the valuable suggestions and comments. Hope this revision has improved the paper to a level of your satisfaction. Response to your comments are given below sequentially.

 

Comment # 1: How does this compare with state-of-the-art commercial and experimental software packages with similar functionality?

Response: Authors are thankful to the respected reviewer for valuable comment. The response of the authors is following;

The main contribution of the proposed manuscript is the integrated motion planning (for robotic manipulator and parts to be assembled). To the best of the authors knowledge, the integrated motion planning is currently an ongoing research topic. To the best of authors knowledge, there are no commercially available packages specifically for this kind of problems. There are some software packages available such as OMPL (Open Motion Planning Library) which can be integrated with ROS-based platform. However, the OMPL only includes basic planning algorithm such as RRT, PRM etc., and their variants but it does not take into account the constraints of the assembly task. Authors have implemented the proposed algorithm in the manuscript using OMPL. It is mentioned in the last para of section 4 (Experimental Results) of the revised manuscript.

 

Comment # 2: Inclusion of a video to support the snapshots and the results

Response: Authors have provided a short video to demonstrate the 4 pieces SOMA puzzle assembly.

Author Response File: Author Response.pdf