Literature Review

Introduction

The wheel-legged biped robots are a type of hybrid robot that combines the advantages of both wheeled and legged locomotion. The development of Wheel legged robot has gained significant attention in recent years due to its potential to combine the benefits of wheeled robots and legged robots. The robot has a wheel-leg mechanism that allows it to move with speed and agility on flat surfaces and climb over obstacles with the help of legs. [8] This review aims to provide an overview of the current state of the art in wheel-legged biped robots, including their design, control, and applications. A bipedal wheel-legged robot is a type of robot that combines the advantages of both legged and wheeled robots. Legged robots are versatile and can navigate through rough terrain, while wheeled robots are efficient and can easily manoeuvre on flat ground. By combining these two types of locomotion, a bipedal wheel-legged robot can achieve both versatility and efficiency in its movements. [9] However, designing and controlling a bipedal wheel-legged robot is challenging because it is an under-actuated system. An under-actuated system is one that has fewer control inputs than degrees of freedom, meaning that it is difficult to achieve the desired movements and stability. In the case of a bipedal wheel-legged robot, it only has two contact points with the ground, which adds to the complexity of designing its structure and controlling its movements. [10]

There have been many successful research results in both the structure design and control methods of bipedal wheel-legged robots. Researchers have explored different approaches to designing the structure of the robot, such as using a torso to improve stability, or using multiple sets of wheels to increase the contact points with the ground. Additionally, various control methods have been proposed, including model-based control, learning-based control, and hybrid control. These methods enable the robot to achieve stable locomotion and perform tasks with high accuracy. [11] While there are several wheel-legged prototypes on the market that have shown excellent performance, including Handle from Boston Dynamics, Ascento from ETH, Ollie from Tencent, WRL from HIT, and NeZha from SUSTech, these systems have limitations when it comes to recovering from external impact or disturbance with minimal displacement. However, the potential benefits of combining the versatility of legged robots with the efficiency of wheeled robots make the development of bipedal wheel-legged robots a worthwhile pursuit. The successful outcome of the balance control applied to an actual Wheel legged robot platform has been performed in the wheel legged robot. The researchers developed a control algorithm that integrates the dynamic equations of motion and feedback control. The algorithm was implemented on a prototype in wheel legged robot platform, and the results showed that the platform was able to maintain its balance while moving on a slope. [12] [13]

Wheeled robots have become popular in the robot industry due to their high efficiency in execution. [14] However, they have limitations in adapting to complex terrains. On the other hand, legged robots, especially biped robots, have a long history of research and are more adaptable to complex environments. Biped robots have developed to the level of humanoid robots, which can adapt to human working environments and solve many problems effectively. However, biped robots require complex control algorithms to maintain dynamic stability. The advantages of wheeled robots and biped robots are complementary to each other. Therefore, the concept of the wheeled legged robot (WLR) has emerged. WLR has the strong terrain adaptability of legged robots and the fast response and execution accuracy of wheeled robots. [15] If the problem of motion balance can be well solved and artificial intelligence technology is used as the brain, WLR will have broad application scenarios and great commercial value. WLR has high research value due to its strong adaptability and manoeuvrability. By combining the advantages of wheeled robots and biped robots, WLR can potentially solve some of the limitations of both and become a versatile and efficient robot for various applications. Biped robots have been studied extensively as an important sub-class of legged robots, with the aim of understanding human locomotion and manipulation and making them better serve humans in living environments. [16] [17]

Applications

Wheel-legged biped robots have a wide range of applications, including search and rescue, military, and entertainment. In search and rescue operations, the robot can traverse rough terrain and access hard-to-reach areas to search for survivors. In the military, the robot can be used for reconnaissance and surveillance missions. In entertainment, the robot can be used for theme park rides or as a companion robot. [16] The challenges posed by their under-actuated nature and limited contact points with the ground require innovative solutions in both structure design and control methods. In search and rescue, these robots can be used to access difficult-to-reach areas, such as collapsed buildings or disaster zones, where wheeled or legged robots may not be suitable. In military applications, these robots can provide mobility and agility in challenging terrains, such as mountains or jungles. [18]

Also wheel legged robot has several potential application scenarios, such as in airports, hotels, and pension service industries. It can also be used as teaching aids, entertainment films, and other aspects in colleges and universities. In entertainment, these robots can be used for performing stunts or creating special effects in movies or shows. [19]

Existing Robots

One of the key challenges in controlling wheeled bipedal robots is maintaining balance. The Ascento 2 robot uses a combination of feedback and feedforward control to achieve balance. The feedback control system constantly monitors the robot's posture and makes adjustments to maintain stability. The feedforward control system uses predictive models of the robot's motion to anticipate changes in the robot's posture and make pre-emptive adjustments to maintain stability. Another challenge in controlling wheeled bipedal robots is achieving efficient locomotion. The Ascento 2 robot uses a gait pattern that alternates between using its legs and wheels for propulsion, depending on the terrain. This allows the robot to move quickly and efficiently over a wide range of terrain. [1]

Wheeled bipedal robots offer a promising approach to mobile ground robot deployment, combining the efficiency and speed of wheels with the strength and versatility of legs. Achieving balance and efficient locomotion in these systems requires advanced control techniques that take into account the nonlinearity and hybrid nature of the dynamics. The Ascento 2 robot is a good example of a successful implementation of such a system. [20]

Figure 1. Ascento 2 Robot [1]

The Ascento 2 robot provides a good example of how to control and balance a wheeled bipedal robot. This type of robot combines the advantages of both wheels and legs for efficient and agile movement over varied terrain. The control system must be able to handle the complex dynamics of such a system, and Ascento 2 achieves this by using a combination of feedback and feedforward control to maintain balance and efficient locomotion. The control system must take into account the nonlinearity and hybrid nature of the dynamics, as well as the need for coordination between the wheels and legs. This can be achieved through the use of advanced control techniques such as model predictive control, nonlinear control, and hybrid control. [1]

Ascento 02 Robot

In addition, these systems often use complex control algorithms that are designed to optimize specific performance criteria, such as speed or efficiency. While these algorithms may be effective under normal conditions, they may not be able to adapt to sudden changes in the environment or unexpected disturbances. As a result, bipedal wheel-legged robots often struggle to recover from external impacts or disturbances with minimal displacement, and in some cases, they may fail to recover at all. To address these limitations, researchers are developing new control algorithms and strategies that are better able to handle the complex dynamics of bipedal wheel-legged robots and respond to external disturbances in real-time. [21]

Figure 2.Wheel legged robot prototype [21]

Bipedal wheel-legged robots are unique in that they combine the advantages of both legged robots and wheeled robots. They are able to move effectively on flat terrain like wheeled robots, while also being adaptable to uneven terrain like legged robots. However, they also present greater difficulties due to having only two points of contact with the ground, compared to other types of bipedal robots that have more points of contact. One of the main reasons for this limitation is the internal behaviour of these systems. The dynamics of bipedal wheel-legged robots are complex and highly nonlinear, making them difficult to control. Even small disturbances can cause significant changes in the robot's motion, making it difficult to maintain stability and recover from impacts.

A Biped Wheel Legged Robot

Figure 3. The simulation model in Matlab Simulink with simscapes multibody is shown in A and the corresponding visual display is shown in B. [21]

Before experimental validation, a simulation model is typically constructed to test and refine the design of the robot. In this case, the simulation model is built using MATLAB Simulink Simscape Multibody. To simplify the simulation model, the parallelogram link between the knee motor and knee joint is treated as rigidly linked to the thigh. This is because the displacement of this link is very small, and treating it as rigid simplifies the simulation model. In addition, for ease of controller design, the knee motors and hip motors are operated in position mode, meaning that they are controlled to move to a specific position rather than applying a specific force. The E-Jet, which is a type of actuator, is also simplified in the simulation model to be equivalent to a controllable force applied at the centre of where it is installed. Simplifying the simulation model in this way makes it easier to design and test control algorithms and refine the design of the robot before experimental validation. Once the simulation model is validated, it can be used to optimize the design and control of the robot for improved performance in real-world applications. [21]

A self-balancing wheel-legged robot is a type of robot that has a smaller footprint, can move faster, and is more manoeuvrable than legged biped robots. This is because it uses a differential drive, which allows for non-holonomic movement constraints and self-balancing abilities. However, there are still challenges that need to be addressed in path planning, motion control, stability analysis, and system modelling to improve its performance. Due to its distinctive qualities, self-balancing wheel-legged robots are particularly suitable for indoor environments such as hotel lobbies, banks, hospitals, restaurants, and warehouses. There are two types of wheeled inverted pendulum (WIP) robots: those with legs and those without, and earlier investigations in this field primarily focused on the lower body of the robot. To enhance the stability and robustness of the underactuated self-balancing wheel-legged robot, a control system is presented. Additionally, the impact of using an active arm on the robot's improved balance stability is statistically assessed in the ROS-Gazebo environment using the centroidal moment pivot (CMP) as a critical indicator. The control system is designed to improve the robot's stability and robustness by addressing the challenges in path planning, motion control, stability analysis, and system modelling. This helps to optimize the performance of the robot in real-world environments. Furthermore, the use of an active arm is shown to improve the robot's balance stability, which is measured using the CMP. The CMP is a critical indicator of the robot's balance stability and is used to assess the impact of using an active arm on the robot's performance. Self-balancing wheel-legged robots have distinctive qualities that make them attractive for indoor environments. However, there are still challenges that need to be addressed to optimize their performance. The presented control system and the use of an active arm are examples of approaches to improve the robot's stability and robustness. [22]

Figure 4. Wheel legged robot manipulator arm [22]

Wheel Legged Robot with Manipulator Arm

A time-varying linear quadratic regulator (LQR) is designed for the multi-rigid body system, and a model predictive controller (MPC) is designed for the wheeled inverted pendulum. The time-varying LQR is designed to achieve velocity tracking, while the MPC is designed to achieve changing height, resist external disturbances, and enable jumping. By combining these two control strategies, the WBR is capable of achieving stable dynamic motion and maintaining balance. The effectiveness of the control framework is verified through both physical experiments and simulations. The proposed control strategy is designed to tackle the challenges of controlling an underactuated nonlinear system like a WBR. The decoupled control structure helps to simplify the control problem, and the use of different control strategies for different simplified models allows for the achievement of multiple dynamic behaviours. [23]

The statement describes the advantages of combining leg-based and wheeled locomotion in a single robot. Legged robots can navigate through difficult terrain and can adapt to their environment, while wheeled robots are highly efficient and can move quickly on flat surfaces. By combining the two, a robot can achieve the best of both worlds. The statement also mentions the use of masses and centroid positions to weight the position of the analogous centroid. [23] In engineering, the centroid is the centre of mass of a particular object, which is determined by the distribution of its mass. For a robot, the centroid represents the location of its overall mass. To create a kinematic model, the Denavit-Hartenberg (D-H) convention is used. The D-H convention is a commonly used method to define the relationship between the links and joints of a robot. The convention uses a set of parameters to define the orientation and position of each joint in relation to the previous joint. Using the D-H convention, the relationship between the centroid of the robot's mass and the axle coordinate system can be defined. This allows for precise control of the robot's motion, as the position of the centroid can be used to determine the robot's overall balance and stability. The statement describes the use of both legged and wheeled locomotion in a single robot and the use of engineering principles such as centroid position and the D-H convention to create a kinematic model for precise control of the robot's motion. [24]

Figure 5. Robot Prototype and Simplified model [23]

The statement discusses a control strategy that can enhance the stability and dynamic motion of a wheeled biped robot (WBR), which is a complex, underactuated nonlinear system that is challenging to control. Underactuated systems have fewer actuators than the degrees of freedom of the system, making it difficult to achieve stable control. The proposed control strategy involves decoupling the WBR into two simplified models: a five-link multi-rigid body system and a variable-length wheeled inverted pendulum. This simplification helps to make the control problem more manageable.

Controlling of a Wheel legged biped robot

Additionally, they can interact with their surroundings using their arms and hands, making them well-suited to perform tasks that require manipulation of objects. [25] While legged robots are less efficient in their movement compared to wheeled robots, they are more versatile and can adapt to different types of terrain and environments.

The combination of bipedal and wheeled robots, also known as WLR, offers significant advantages over either type of robot alone. By combining the rapid response and execution accuracy of wheeled robots with the excellent terrain adaptability of legged robots, WLRs can operate in a wide range of environments and perform a variety of tasks with high efficiency and precision. If the problem of motion balance is successfully solved and artificial intelligence technology is deployed as the brain, WLRs will have numerous application scenarios and high economic value. They can be used as instructional aids, entertainment in films, military equipment, and other areas in schools and universities, as well as in airports, hotels, pensions, and other service businesses. Furthermore, because of its superior adaptability and agility, WLR has high research value. Researchers can explore and develop new technologies and applications for WLR, which will enhance its capabilities and further expand its range of applications. [25]

Figure 6. Wheel legged Robot Prototype [25]

Wheeled robots have been extensively used in various applications such as industrial manufacturing, logistics, and military operations. They are highly efficient in their execution of tasks due to their fast mobility and high payload capacity. However, they are not well-suited to adapt to complex terrain and environments. Wheeled robots can get stuck or topple over when faced with obstacles such as stairs, rocks, or uneven surfaces. To address this issue, researchers have been working on developing wheeled robots that can perform a variety of tasks and also navigate complex terrains. However, despite the progress made in this area, wheeled robots are still not comparable to legged robots in terms of terrain adaptability. Legged robots, particularly bipedal robots, have a long history of research on their legs and have advanced to the point where they can function as humanoid robots. Bipedal robots can adapt to the working environment of humans and resolve various issues. They have the ability to traverse complex terrain such as stairs, slopes, and uneven surfaces.

LQR Two-wheel legged robot

Pros and Cons

Considering about the pros, the robots with driven wheels have the potential to move faster and more efficiently than legged locomotion on continuous flat ground. As a result, many researchers have attempted to combine wheels with bipedal robots to overcome some of the shortcomings of humanoid robots. These bipedal leg-wheeled robots have the potential to demonstrate better performance in agility, moving speed, and energy efficiency, making them more suitable for indoor and outdoor inspection and logistics distribution tasks. The combination of legs and wheels in a robot allows it to benefit from the strengths of both types of locomotion. The legs provide the robot with stability and the ability to adapt to uneven terrain, while the wheels provide the robot with speed and efficiency on flat ground. [8] This makes bipedal leg-wheeled robots a promising candidate for various practical applications.

However, biped robots have several inherent disadvantages, such as natural instability, poor flexibility, slow moving speed, and low energy efficiency, which limit their practical applications. [12]

Design

When designing a wheel-legged biped robot in SolidWorks, you would need to take into account the various components of the robot's body, including the base with the wheels, knee joints, shoulder joints. To address the challenge of transitioning between wheel and legged locomotion, you would need to incorporate control strategies into the design. This could involve planning the gait of the robot, which refers to the sequence of limb movements required for locomotion. The gait planning could be done using software Matlab. Another important aspect of the robot's design would be the motion control system. This could involve selecting appropriate motors and actuators for the wheels and legs, and programming them to perform the desired movements.

Additionally, sensory feedback control would also be important in ensuring a smooth transition between the two modes of locomotion. This could involve using sensors such as accelerometers and gyroscopes to provide feedback on the robot's orientation and position, and adjusting the control signals accordingly. The design of a wheel-legged biped robot in SolidWorks would require careful consideration of the various components and control strategies involved in achieving smooth and efficient locomotion.

Control

In order to design a wheel-legged biped robot, it is important to consider the control aspect, which is essential for stable and efficient locomotion. Matlab Simulink is a popular software used to simulate and control robotic systems. In Simulink, the Simscape Multibody Toolbox can be used to model and simulate the dynamics of the WLR system.

One approach to control the WLR is the model-based control method, which relies on accurate mathematical models of the WLR's dynamics to control its motion. This method is computationally expensive and requires accurate modelling of the WLR's dynamics. Another approach is the learning-based control method, which uses machine learning algorithms to learn the WLR's dynamics and control its motion. This method is computationally efficient and can adapt to changes in the environment.

A hybrid control method can combine both model-based and learning-based control strategies to achieve stable and efficient locomotion. In this case, Simulink can be used to implement the hybrid control approach for the WLR. The balance control algorithm can be implemented in Simulink using feedback control to adjust the torque on the wheels and legs to maintain the balance of the WLR. The dynamic equations of motion can be used to describe the relationship between the external forces acting on the WLR and the internal forces generated by the WLR's wheels and legs.

The use of Simulink and Simscape Multibody Toolbox allows for efficient simulation and control of the WLR system. By simulating the WLR's behaviour in Simulink, it is possible to optimize the control algorithms and evaluate the performance of the WLR under different scenarios and terrains.

Conclusion

Wheel-legged biped robots are a unique type of hybrid robot that possess the advantages of both wheeled and legged locomotion. They are designed to move on both wheels and legs, which allows them to tackle a variety of terrains that would be challenging for robots that only have wheeled or legged locomotion. [1] The design of these robots is a complex task that involves integrating the mechanisms of both wheels and legs. The robot's body typically consists of a base with two wheels.

The wheel-leg mechanism allows the robot to switch between rolling on its wheels and walking on its legs, depending on the terrain. One of the significant challenges in designing such robots is the control of the transition between wheel and legged locomotion. Researchers have proposed various control strategies, including gait planning, motion control, and sensory feedback control, to achieve a smooth transition between the two modes of locomotion. The control of these robots is another challenging aspect that researchers have been working on. Various control strategies, including model-based control, learning-based control, and hybrid control, have been proposed to achieve stable and efficient locomotion. [12]

The balance control algorithm uses feedback control to adjust the torque on the wheels and legs to maintain the balance of the WLR. The dynamic equations of motion are used to describe the relationship between the external forces acting on the WLR and the internal forces generated by the WLR's wheels and legs. The combination of these two approaches allows the WLR to maintain its balance on different types of terrains. Matlab Simulink is a powerful software tool that can be used to simulate the behaviour of wheel-legged biped robots. Simulating the robot's behaviour using Simulink can help researchers and designers to understand the robot's performance and optimize its design and control. [18]