Quick Navigation
Purpose:
Land movement is critical to our mission of helping in hazardous areas as it enables the robot to move without a handler entering the hazardous area.
Design Objective:
The goal for this subsystem in the final project is to provide the capability to traverse rough, highly variable terrain. Rather than designing it to function in any specific environment or conditions, we have the aspiration of making the system robust enough to address as many obstacles and scenarios as possible.
Possible designs include the following:
Legs
- Bipedal
- How do biped robots walk?
- “Linear, inverted pendulum model” (Kajita et al., 2010)
- “Zero-tilting moment point (ZMP)” (Kajita et al., 2003)
- Tripedal
-
- STriDER: Self-excited Tripedal Dynamic Experimental Robot
- Unique walking motion using a tripod leg arrangement with a biomimetic walking motion
- Utilizes a strategy of “passive dynamic locomotion” (Ren et al., 2007)
- Introduces the concept of highly energy efficient motion
-
- Quadrupedal
- Learning Quadrupedal Locomotion over Challenging Terrain
- The level of robustness that can be achieved using a quadrupedal design seems to be extremely promising, but the challenge involved in arriving at such a design would likely be just as formidable
- Pentapedal
- Any direction can be forward
- Limbs can be used to accomplish tasks while maintaining three points of contact for more robust balance
Generally, there is the opportunity to choose between a biomimetic design or control system, and a completely original movement technique not inspired by nature. Additionally, these designs listed above generally will likely need to utilize some manner of machine learning to “teach” the robot to develop and refine its gait. This involves a large amount of simulation and translation to real-life prototype testing on variable terrain. Proprioception is also often utilized to help develop the machine’s motion. This archetype of movement system will also be inseparable from the critical balancing function. (Lee et al., 2020)
Wheels
- Tracks and rollers
- Individual wheels
Hybrid
- Virginia Tech: IMPASS robot
- Specifically addresses the problem of dealing with obstacles that are proportionally large when compared to the robot
- Physical dynamics significantly less complex than “walking” locomotion (Hong et al., 2009)
- Decreased need for controller complexity
Avoiding the un-sweet spot
- The un-sweet spot is the size of objects that pose a difficulty to traverse e.g cars can go over small rocks(stones), and very large rocks (Mountains), but rocks the same size as a car are very difficult to traverse
- Reducing the area of this ‘spot’ for our robot is a priority
MVP Literature Review:
The functionality of locomotion in rugged terrestrial environments, including the capabilities of Walking, Crawling, and Climbing, is a key characteristic of most amphibious robots and a growing area of research (Ahmad et al., 2020; Li et al., 2018). However, independent locomotion mechanisms have limited performance abilities. Thus, hybrid locomotion is growing increasingly common, with explorations into wheel-leg, wheel-track, and wheel-leg-track designs to increase functionality and efficiency.
Legged Robots are often tailored to navigate through rugged environments with uneven terrain. The categorization of legged robots is conducted by considering the number of legs used for locomotion on rugged earth environments (Ahmad et al., 2020). Examples include BIPED robots, which are less stable and specifically designed to utilize lift force to overcome the weight of the robot’s body, QUADRUPED robots, which are stable and utilized for applications that have specifications for high payload capacity, flexibility, adaptability, and HEXAPOD robots, which present a strong ability to walk over unstructured terrain (Ahmad et al., 2020).
While legs as a locomotion mechanism presents an element of uniqueness to the robot’s design, other studied designs including Wheels and Tracks are significantly more common (Ahmad et al., 2020). As a well-investigated system, wheels increase the robot’s movement speed across a multitude of terrains. Today’s amphibious robots use wheels as a means of crawling on rugged terrain and transitioning between land and water environments. Similarly, tracks are advantageous in balancing and adaptability, as they are employed to maneuver across challenging terrain with varying inclination angles and are comparatively superior in complex terrain navigation relative to wheels, despite costs to speed (Ahmad et al., 2020).
The depth of research done into the above locomotive designs is substantive; however, they are predominantly used independently. That is, the designs are used either in land or water environments, rarely in both. In addition to the flexibility and range of operation achievable by the design, many pre-existing designs of similar nature trend towards a focus on autonomy. While this has certain benefits, it also tends to limit potential design choices that may favor a human operator. The Packbot and Andros VI are both autonomous vehicles whose designs favor remote operation (Helmick, n.d.; Wang, 2007). Thus, both designs sacrifice data input that may be extremely useful to a human-operator. Specifically, sensor visibility for robots of this design typically only extend to their immediate surroundings and permit only close-to-ground sight through video feedback (Wang, 2007).
While Weldong Wang, Zhijiang Du, and Lining Sun’s tracked robot with a sensor unit on an arm, and other similar designs, offer a solution to the aforementioned problem, there are still improvements to be made (Yamauchi, n.d.). Particularly, there exist no available examples of designs that combine a dynamic range of transportation configurations and a satisfactory array of feedback sensors for user-driven operation.
Among many of the previously mentioned tracked or split-tracked designs, there exists a gap in the versatility of the designs. Multiple designs studied, including the Andros VI, focus solely on climbing stairs or similar levels of specification (Helmick, n.d.; Wang, 2007; Yamauchi, n.d. Lee, 2005). While it would be unreasonable to expect a single design to perform well in any number or type of environment, the ability to account for multiple environments through differing configurations greatly expands the use and versatility of the robot.
After identifying the various gaps in current frontier research, the team has endeavored to design a modular solution that prioritizes environmental adaptability and mission objective flexibility. Research has shown that terrestrial locomotion tends to fall in one of two camps: either legged systems and other similar biomimetic systems versus wheeled or tracked systems. Legged designs typically offer superior flexibility, enabling the robot to maneuver up or down complex terrains, sometimes including the ability to climb. However, these gains are made at the cost of speed. By comparison, tracks and wheels feature the opposite problem, tending to favor more standardized terrains but possessing increased speed. As a result, the team will design a legged system that uses wheels and tracks as its end effectors. The motivation for this decision is that by incorporating the primary elements of either class of locomotion, the advantages of each class can be maximized while the disadvantages can be reduced. When the robot is travelling through environments of low to medium obstacle density, such as roads, gravel paths, or forested trails, the legs will lock in a specified position and the wheels or tracks will be activated to propel the robot. In more complex terrains, such as rubble filled areas, tide pools, or mountain slopes, the wheels/tracks will lock, acting as feet, while the legs move to enable the robot to clamber over or climb larger obstacles.
Beyond hybridizing legged and wheel/track systems, the team has explicitly decided to design the robot such that the wheels can be converted to tracks and vice versa by the operator before the robot is deployed. This would allow operators to use and maintain only one robot that can fulfill multiple purposes. For example, if the operator expects to deploy the robot in swampy marshes or sandy beaches, tracks may be a more suitable option for increased surface area weight distribution. On the other hand, if the operator intends to use this robot in suburban areas or national parks, wheels may be more favorable for their increased speed.
Metrics
Constraints
Relevant Codes and Standards
- Tractive Force Evaluation For Walking Robots
- Dynamic Locomotion With Four And Six-Legged Robots
- Robots and robotic devices – Safety requirements for personal care robots
- Performance evaluation methods of mobile household robots
Next Steps
In order to make progress on this essential functionality, we must first come to a decision regarding the locomotion mechanism we will employ for our robot. This will involve yet another round of concept generation followed by consideration of those concepts and how they will provide the robot with the desired functionalities we have laid out for ourselves.
Developing the Design Space
As the project has progressed, the team has more clearly defined the bounds of our MVP. After intense deliberation, we settled on a hybridized design for land movement. This portion of the robot would be handled by two sub-teams: the Legs sub-team and the Tracks/Wheels sub-team. Taking inspiration from Hyundai’s TIGER we are developing a legged robot with replaceable end effectors (Hyundai Motor Company, 2021). Currently, the end effectors will be either wheels or tracks. This design will enable the robot to traverse multiple types of terrestrial environments, effectively folding in the Climb & Scale Obstacles Function into the Land Movement function. Each sub-team has undergone a cyclic process of concept generation, selection, and refinement as part of their design process. Learn more about this process for each sub-team below:
Citations
- Ahmad, S., & Razib, M. A. (2020, December 21). Review of Locomotion Strategies for Amphibious Robots – A Review by Mohammad Rafeeq, Siti Fauziah Toha. IEEE Access.
- Caron, S. (2018, October 26). How do biped robots walk?. Retrieved from https://scaron.info/robotics/how-do-biped-robots-walk.html
- Engineering ToolBox. (2004). Friction – friction coefficients and calculator. Retrieved from https://www.engineeringtoolbox.com/friction-coefficients-d_778.html
- Helmick, D., et. all. Multi-Sensor, High Speed Autonomous Stair Climbing.
- Hong, D., Jeans, J. B., & Ren, P. (2009). Experimental verification of the walking and turning gaits for a two-actuated spoke wheel robot. Proceedings of the 2009 IEEE/RSJ International Conference on Intelligent Robots and Systems, 402–403.
- Hong, D. [Dennis Hong]. (2007, September 2). STriDER: Self-excited Tripedal Dynamic Experimental Robot [Video]. YouTube. https://www.youtube.com/watch?v=7XsaJwKKBYo
- Hyundai Motor Company. (2021, February 10). Hyundai Motor Group Unveils tiger uncrewed ultimate mobility vehicle concept. Hyundai. Retrieved October 20, 2021, from https://www.hyundai.com/au/en/hyundai-info/news/2021/02/hyundai-motor-group-unveils-tiger-uncrewed-ultimate-mobility-vehicle-concept
- Kajita, S., Kanehiro, F., Kaneko, K., Fujiwara, K., Harada, K., Yokoi, K., & Hirukawa, H. (2003). Biped walking pattern generation by using preview control of zero-moment point. 2003 IEEE International Conference on Robotics and Automation (Cat. No.03CH37422), 1620–1626. https://doi.org/10.1109/ROBOT.2003.1241826
- Kajita, S., Morisawa, M., Miura, K., Nakaoka, S., Harada, K., Kaneko, K., Kanehiro, F., & Yokoi, K. (2010). Biped walking stabilization based on linear inverted pendulum tracking. 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems, 4489–4496. https://doi.org/10.1109/IROS.2010.5651082
- Lee, J., Hwangbo, J., Wellhausen, L., Koltun, V., & Hutter, M. (2020a). Learning Quadrupedal Locomotion over Challenging Terrain. Science Robotics, 5(47). https://doi.org/10.1126/scirobotics.abc5986
- Lee, J. [Robotic Systems Lab: Legged Robotics at ETH Zürich]. (2020b, October 21). Learning Quadrupedal Locomotion over Challenging Terrain [Video]. YouTube. https://www.youtube.com/watch?v=9j2a1oAHDL8
- Lee, W., et. all. (2005, April). Rough Terrain Negotiation Mobile Platform with Passively Adaptive Double-Tracks and Its Application to Rescue Missions. IEEE Access.
- Li, T., & Wang, M. (2018). Review of A Survey on Amphibious Robots by Ziyi Guo. IEEE Access.
- Ren, P., Morazzani, I., & Hong, D. (2007). Forward and Inverse Displacement Analysis of a Novel Three-Legged Mobile Robot Based on the Kinematics of In-Parallel Manipulators. Volume 8: 31st Mechanisms and Robotics Conference, Parts A and B, 1041–1052. https://doi.org/10.1115/DETC2007-34606
- Virginia Tech. (2009, June 26). Virginia Tech: IMPASS robot [Video]. YouTube. https://www.youtube.com/watch?v=16w7i41f7hA
- Wang, W., et. all. (2007, May 8). Dynamic Load Effect on Tracked Robot Obstacle Performance.
- Yamauchi, B. PackBot: A Versatile Platform for Military Robotics.