Fundamento matemático y procedimiento
II.5. Aproximaciones para la difusión intrapartícula
Figures 19 through 24 show a comparson between the simulation air-walking and the real life
air-walking. The process is leaning towards the robot’s right, taking a first step with the left foot,
leaning towards the left, taking a second step with the right foot, and leaning a third time torwards
the right. This then repeats starting with stepping with the left foot.
(a) (b)
(a) (b)
Figure 20: Comparison of TigerBot while leaning towards the right. (a) Simulation (b) Real robot
(a) (b)
(a) (b)
Figure 22: Comparison of TigerBot leaning to the left and forward. (a) Simulation (b) Real robot
(a) (b)
(a) (b)
Figure 24: Comparison of TigerBot leaning to the right and forward. (a) Simulation (b) Real robot
The comparisons match fairly well. If the robot ever needs to have the joint positions calibrated,
the process is moving TigerBot to stand completely straight and zeroing out the encoders so they
are set at zero degrees at the proper spot. The timing is somewhat slower than ideal, however.
Initially, when the timing is slower, the execution is on pace with the desired timing, but when the
Figure 25: The timing diagrams of ROS messages during air-walking with the real robot. The time line starts when the first Teensy sent a message Each Teensy is labeled 0-4, and sends a message with their id when the joints they control are at their desired positions. When all Teensys are done, a message to move to the next pose is sent to each of them, which is put at a value of 5 in the graph. The Done Messages series shows when each message was sent; the y-axis value is shows where the latest message came from. The Theoretical Timing series shows when the ‘move to next position’ messages should be published.
The executed air-walk was short of the theoretical timing by two seconds, taking about 16
seconds instead of about 14 seconds. This could be fixed by increasing the speeds the joints move
at so they reach their target positions sooner.
Chapter 7: Conclusion
TigerBot is a platform for RIT for graduate students to research humanoid robots. At full-scale,
it is more useful than smaller-scale robot as it can walk and perform human-like motion with more
fidelity. The work done on TigerBot so far has been to set up both hardware and software to control
the real robot and also to run simulations. This involved integrating the encoders and motors with
ROS so they can be controlled with the ODroid, and creating a model and setting up the simulator.
To test both, a walking gait is developed and tested on the simulation and the real robot. The
simulated robot walked while the real robot went through the motions while suspended. The 0 1 2 3 4 5 0 2 4 6 8 10 12 14 16 Com p le tio n m es sage ID Time (s)
Timing Diagram of Air-Walk Execution
success in both the simulation and the air-walking acts as proof of concept and leaves the robot
ready to be used by other students as is or to be improved upon further.
Chapter 8: Future Work
Additional work can be done to the real robot. The 2 DOF arms on the model have been
designed but not machined and attached. The currently unused sensors, including the current
sensors, force sensors on the feet, and the IMU require code to integrate with the single board
computer. A camera could also be attached as a head for computer vision, such as a RBG-D camera
like the Kinect or Xtion Pro.
With the physical platform fully set up, more robust walking could be implemented following
the demonstrated walking gait with modifications. For example, the force sensors on the feet could
be used to determine the best gait for walking on rough terrain, or how to recover from being
pushed.
To move towards TigerBot’s original premise of being a tour guide, a camera as mentioned
earlier or GPS could be used for navigation while walking, to perform Simultaneous Localization
and Mapping (SLAM) within a building or to navigate to specific coordinates when outside. This
would also require changing the power supply to batteries, as TigerBot is currently tethered. The
batteries would need to be strong enough for TigerBot to walk a significant distance and would
likely be heavy enough to affect TigerBot’s mass. The batteries could be pulled along behind
TigerBot on a cart, but it would be more conventional to carry the batteries in its back. Either way,
to be able to simulate walking with such modifications, the URDF would have to be edited, either
Implementing the simulated walking gait on the real robot requires IMU data for the PI
controller using COM velocity, with the data adjusted to estimate the hips’ velocity instead of the torso’s, and calculating the location of the feet relative to the hips. Running the PI controller on the ODroid with the same conditions as it had with the simulation will output changes to the ankles’
flex joint angles and the timing of the positions. These changes would need to be published so that
the node that controls the timing and the Teensys that control the ankle see them and adjust
accordingly.
Should future users want to continue with the walking gait developed in simulation, it would
be best to focus on improving the gait’s speed. This would require more stabilization, so perhaps a different method of balancing besides the PI controller could be explored. Otherwise, different
walking methods, such as modeling TigerBot as three masses or involving the active toe joints,
Bibliography
[1] Hirai, K., M. Hirose, Y. Haikawa, and T. Takenaka. “The Development of Honda Humanoid Robot.” In Proceedings. 1998 IEEE International Conference on Robotics and Automation (Cat. No.98CH36146), 2:1321–26 vol.2, 1998. https://doi.org/10.1109/ROBOT.1998.677288.
[2] Sakagami, Y., R. Watanabe, C. Aoyama, S. Matsunaga, N. Higaki, and K. Fujimura. “The Intelligent ASIMO: System Overview and Integration.” In IEEE/RSJ International Conference on Intelligent Robots and Systems, 3:2478–83 vol.3, 2002.
https://doi.org/10.1109/IRDS.2002.1041641.
[3] Kaneko, K., F. Kanehiro, S. Kajita, K. Yokoyama, K. Akachi, T. Kawasaki, S. Ota, and T. Isozumi. “Design of Prototype Humanoid Robotics Platform for HRP.” In IEEE/RSJ International Conference on Intelligent Robots and Systems, 3:2431–36 vol.3, 2002.
https://doi.org/10.1109/IRDS.2002.1041632.
[4] “System Design and Dynamic Walking of Humanoid Robot KHR-2.” In Proceedings of the 2005 IEEE International Conference on Robotics and Automation, 1431–36, 2005.
https://doi.org/10.1109/ROBOT.2005.1570316.
[5] Gouaillier, D., V. Hugel, P. Blazevic, C. Kilner, J. Monceaux, P. Lafourcade, B. Marnier, J. Serre, and B. Maisonnier. “Mechatronic Design of NAO Humanoid.” In 2009 IEEE International Conference on Robotics and Automation, 769–74, 2009.
https://doi.org/10.1109/ROBOT.2009.5152516.
[6] T. Buschmann, and H. Ulbrich. “Humanoid Robot LOLA.” In 2009 IEEE International Conference on Robotics and Automation, 775–80, 2009.
https://doi.org/10.1109/ROBOT.2009.5152578.
[7] V. Favot, S. Lohmeier, M. Schwienbacher, and H. Ulbrich. “Experiments in Fast Biped Walking.” In 2011 IEEE International Conference on Mechatronics, 863–68, 2011.
https://doi.org/10.1109/ICMECH.2011.5971235.
[8] Wahrmann, D., A. Hildebrandt, R. Wittmann, F. Sygulla, D. Rixen, and T. Buschmann. “Fast Object Approximation for Real-Time 3D Obstacle Avoidance with Biped Robots.” In 2016 IEEE International Conference on Advanced Intelligent Mechatronics (AIM), 38–45, 2016.
https://doi.org/10.1109/AIM.2016.7576740.
[9] A. Hildebrandt, D. Wahrmann, F. Sygulla, D. Rixen, and T. Buschmann. “Model-Based Predictive Bipedal Walking Stabilization.” In 2016 IEEE-RAS 16th International Conference on Humanoid Robots (Humanoids), 718–24, 2016. https://doi.org/10.1109/HUMANOIDS.2016.7803353. [10] T. Buschmann, and H. Ulbrich. “System Design and Control of Anthropomorphic Walking
Robot Lt;Emphasis Emphasistype="italic" Gt;LOLA Lt;/Emphasis Gt;” IEEE/ASME Transactions on Mechatronics 14, no. 6 (December 2009): 658–66.
https://doi.org/10.1109/TMECH.2009.2032079.
[11] S. Lohmeier, M. Bachmayer, H. Ulbrich, and F. Pfeiffer. “A Collocation Method for Real-Time Walking Pattern Generation.” In 2007 7th IEEE-RAS International Conference on Humanoid Robots, 1–6, 2007. https://doi.org/10.1109/ICHR.2007.4813841.
[12] Vukobratovic, Miomir, and Branislav Borovac. “Zero-Moment Point - Thirty Five Years of Its Life.” I. J. Humanoid Robotics 1 (March 1, 2004): 157–73.
https://doi.org/10.1142/S0219843604000083.
[13] Vukobratovic, M., and D. Juricic. “Contribution to the Synthesis of Biped Gait.” IEEE Transactions on Biomedical Engineering BME-16, no. 1 (January 1969): 1–6.
[14] Kajita, Shuuji. Introduction to Humanoid Robotics. 1st edition. Springer Tracts in Advanced Robotics 101. New York: Springer, 2014.
[15] Hugel, V., and N. Jouandeau. “Walking Patterns for Real Time Path Planning Simulation of Humanoids.” In 2012 IEEE RO-MAN: The 21st IEEE International Symposium on Robot and Human Interactive Communication, 424–30, 2012.
https://doi.org/10.1109/ROMAN.2012.6343789.
[16] Shibuya, M., T. Suzuki, and K. Ohnishi. “Trajectory Planning of Biped Robot Using Linear Pendulum Mode for Double Support Phase.” In IECON 2006 - 32nd Annual Conference on IEEE Industrial Electronics, 4094–99, 2006. https://doi.org/10.1109/IECON.2006.348126. [17] Tsuji, T., and K. Ohnishi. “A Control of Biped Robot Which Applies Inverted Pendulum Mode
with Virtual Supporting Point.” In 7th International Workshop on Advanced Motion Control. Proceedings (Cat. No.02TH8623), 478–83, 2002. https://doi.org/10.1109/AMC.2002.1026967. [18] Park, J., and Y. Youm. “General ZMP Preview Control for Bipedal Walking.” In Proceedings
2007 IEEE International Conference on Robotics and Automation, 2682–87, 2007.
https://doi.org/10.1109/ROBOT.2007.363870.
[19] Missura, M., and S. Behnke. “Online Learning of Foot Placement for Balanced Bipedal Walking.” In 2014 IEEE-RAS International Conference on Humanoid Robots, 322–28, 2014.
https://doi.org/10.1109/HUMANOIDS.2014.7041379.
[20] Luo, D., Y. Wang, and X. Wu. “Online Learning of COM Trajectory for Humanoid Robot Locomotion.” In 2012 IEEE International Conference on Mechatronics and Automation, 1996– 2001, 2012. https://doi.org/10.1109/ICMA.2012.6285128.
[21] Ames, Aaron D. “First Steps toward Automatically Generating Bipedal Robotic Walking from Human Data.” In Robot Motion and Control 2011, 89–116. Lecture Notes in Control and Information Sciences. Springer, London, 2012. https://doi.org/10.1007/978-1-4471-2343-9_8. [22] Ames, A. D. “Human-Inspired Control of Bipedal Walking Robots.” IEEE Transactions on
Automatic Control 59, no. 5 (May 2014): 1115–30. https://doi.org/10.1109/TAC.2014.2299342. [23] Chen, X., Y. Zhou, Q. Huang, Z. Yu, G. Ma, L. Meng, and C. Fu. “Bipedal Walking with Toe-
off, Heel-Strike and Compliance with External Disturbances.” In 2014 IEEE-RAS International Conference on Humanoid Robots, 506–11, 2014.
https://doi.org/10.1109/HUMANOIDS.2014.7041409.
[24] Li, Z., B. Vanderborght, N. G. Tsagarakis, and D. G. Caldwell. “Human-like Walking with Straightened Knees, Toe-off and Heel-Strike for the Humanoid Robot ICub.” In UKACC International Conference on Control 2010, 1–6, 2010. https://doi.org/10.1049/ic.2010.0356. [25] You, Y., S. Xin, C. Zhou, and N. Tsagarakis. “Straight Leg Walking Strategy for Torque-
Controlled Humanoid Robots.” In 2016 IEEE International Conference on Robotics and Biomimetics (ROBIO), 2014–19, 2016. https://doi.org/10.1109/ROBIO.2016.7866625. [26] Huang, Qiang, K. Yokoi, S. Kajita, K. Kaneko, H. Arai, N. Koyachi, and K. Tanie. “Planning
Walking Patterns for a Biped Robot.” IEEE Transactions on Robotics and Automation 17, no. 3 (June 2001): 280–89. https://doi.org/10.1109/70.938385.
[27] Ogura, Y., K. Shimomura, H. Kondo, A. Morishima, T. Okubo, S. Momoki, H. Lim, and A. Takanishi. “Human-like Walking with Knee Stretched, Heel-Contact and Toe-off Motion by a Humanoid Robot.” In 2006 IEEE/RSJ International Conference on Intelligent Robots and Systems, 3976–81, 2006. https://doi.org/10.1109/IROS.2006.281834.
[28] Erbatur, K., and O. Kurt. “Natural ZMP Trajectories for Biped Robot Reference Generation.” IEEE Transactions on Industrial Electronics 56, no. 3 (March 2009): 835–45.
https://doi.org/10.1109/TIE.2008.2005150.
[29] Yi, S., and D. D. Lee. “Dynamic Heel-Strike Toe-off Walking Controller for Full-Size Modular Humanoid Robots.” In 2016 IEEE-RAS 16th International Conference on Humanoid Robots (Humanoids), 395–400, 2016. https://doi.org/10.1109/HUMANOIDS.2016.7803306.
[30] Park, J. H., and K. D. Kim. “Biped Robot Walking Using Gravity-Compensated Inverted Pendulum Mode and Computed Torque Control.” In Proceedings. 1998 IEEE International Conference on Robotics and Automation (Cat. No.98CH36146), 4:3528–33 vol.4, 1998.
https://doi.org/10.1109/ROBOT.1998.680985.
[31] Shimmyo, S., T. Sato, and K. Ohnishi. “Biped Walking Pattern Generation by Using Preview Control Based on Three-Mass Model.” IEEE Transactions on Industrial Electronics 60, no. 11 (November 2013): 5137–47. https://doi.org/10.1109/TIE.2012.2221111.
[32] Sato, T., S. Sakaino, and K. Ohnishi. “Real-Time Walking Trajectory Generation Method With Three-Mass Models at Constant Body Height for Three-Dimensional Biped Robots.” IEEE Transactions on Industrial Electronics 58, no. 2 (February 2011): 376–83.
https://doi.org/10.1109/TIE.2010.2052535.
[33] Luo, R. C., and C. C. Chen. 2018. “Quasi-Natural Humanoid Robot Walking Trajectory Generator Based on Five-Mass With Angular Momentum Model.” IEEE Transactions on Industrial Electronics 65 (4): 3355–64. https://doi.org/10.1109/TIE.2017.2750628.