[Abstract] This paper focuses on the research and development of robot mobility functions, developing various mobility mechanisms to adapt to different requirements in various working environments. Among them, omnidirectional wheels can achieve high-precision positioning, on-the-spot posture adjustment, and movement along arbitrary continuous trajectories on a two-dimensional plane, possessing unique characteristics that cannot be replaced by general wheeled mobility mechanisms. This is of great significance for the study of free walking of mobile robots. The paper comprehensively analyzes the existing mobility mechanisms and movement characteristics of mobile robots, and analyzes their kinematic features. A stability determination method for static walking of mobile robots is proposed, and the robot's straight-line walking gait, fixed-point turning gait, and obstacle-crossing walking gait are planned. A PLC is used to control the gait.
introduction
A mobile robot is a robotic system capable of sensing its external environment and its own state through sensors, enabling autonomous movement towards a target in obstacle-filled environments to complete specific tasks. In recent years, the increasingly broad application prospects of mobile robots in industry, agriculture, medicine, aerospace, and various aspects of human life have made them a research hotspot in international robotics. Since the 1990s, research on mobile robots has reached a higher level, marked by the development of advanced environmental information sensors and information processing technologies, highly adaptable mobile robot control technologies, and planning technologies for real-world environments. Currently, mobile robots, especially autonomous robots, have become a highly active research area in robotics technology.
From the earliest robots to the diverse mobile vehicles that have emerged today, the forms of their mobility mechanisms have proliferated. Developed Western countries, led by the United States, Russia, France, and Japan, have developed a variety of complex and unique three-dimensional mobility mechanisms, some of which have already entered the practical and commercialization stages. Facing the challenges of deep space exploration in the 21st century, the development of various autonomous systems is essential, and mobility mechanisms are the most basic and crucial component of these systems. The mobility mechanisms of existing mobile robots mainly include tracked, legged, and wheeled types. Wheeled mechanisms are the most efficient but have relatively poor adaptability, while legged mechanisms have the strongest adaptability but the lowest efficiency. Tracked mobility mechanisms consist of a circular track wound around several wheels, preventing the wheels from directly contacting the ground. The tracks help to mitigate uneven terrain. They offer good stability, obstacle-crossing ability, and a long service life, making them suitable for traversing rugged terrain. However, the heavy tracks and numerous drive wheels make the overall mechanism bulky and consume relatively high power.
Wheeled mobile mechanisms offer advantages such as high speed, high energy efficiency, simple structure, convenient control, and the ability to leverage mature automotive technologies. However, their off-road performance is relatively weak. But with the emergence of various wheeled chassis, such as NASDA's six-wheeled flexible chassis lunar rover LRTV, TRANSMASH's six-wheeled three-body flexible frame mobile robot Marsokohod, CMU's six-wheeled three-body flexible robot Robby series, and JPL's six-wheeled rocker-suspended planetary rover Rocky series, the off-road capabilities of wheeled robots have been greatly enhanced, rivaling those of legged robots. Consequently, research on robot mechanisms has shifted to wheeled mechanisms, especially with the recent development in Japan of the unique five-point support suspension structure Micros, whose superior off-road capabilities surpass even those of legged robots.
1. Selection of robot movement mode and structural design
1.1 Selection of Movement Mode
Currently, the mainstream modes of transportation are wheeled, legged, and tracked. However, due to their respective advantages and disadvantages, scientists are increasingly pursuing improvements in overall performance. Wheeled mobility mechanisms offer advantages such as high speed, high energy efficiency, simple structure, convenient control, and the ability to draw upon mature automotive technologies; however, their off-road performance is relatively weak. Legged mobility structures, while possessing excellent off-road capabilities, suffer from complex structures and low efficiency. Tracked mobility mechanisms are cumbersome due to the heavy tracks and numerous drive wheels, resulting in a relatively large power consumption.
The design environment for this project is primarily man-made, with relatively flat terrain, but obstacles such as steps and stairs still need to be considered. Therefore, I plan to design a combined wheel-leg mobility system. On flat roads, the wheeled structure offers high efficiency and speed, while the legged structure is used to overcome obstacles such as steps and stairs. Since the robot includes a legged structure and needs to climb steps and stairs, a four-legged structure is adopted. This is because while a wheeled structure can handle steps, it cannot handle stairs, hence the need for a legged structure. Stairs are ubiquitous in daily life, and overcoming the challenge of climbing stairs is essential for the robot's adaptability. I decided on a four-wheeled legged structure, and the basic structure is shown in Figure 1. The central part is the robot's main body, containing the control system and motors that drive the upper limbs. The four limbs end in tires. Each leg is divided into an upper limb and a lower limb, with a joint in the middle, around which the lower limb can rotate.
Figure 1 Basic structure of the robot
1.2 Conceptualization of Robot Movement Principle
Because the environment is relatively good, mostly flat, the primary mode of movement is wheeled mobility. Legged structures are only used when climbing stairs or steps, as these are less efficient. Using legged structures only when absolutely necessary improves the robot's movement efficiency and provides better obstacle-crossing capabilities. The principles for climbing stairs and steps are essentially the same, so I will only explain my concept for the stair-climbing movement principle.
First, sensors need to be installed on the robot's body so that it can detect obstacles such as stairs ahead. Then comes the process of climbing the stairs. Before climbing, the braking system on the wheels must be activated to prevent the wheels from turning. Then, the climbing process is like a person walking up stairs: the robot takes turns stepping on the front foot, and once the front foot is stable, it takes turns stepping on the back foot.
1.3 Selection of Robot Wheels
There are many types of wheels on the market now, including standard wheels, casters, Mecanum wheels, spherical wheels, and orthogonal wheels. I decided to choose Mecanum wheels because they can move well in all directions and are not as difficult to control as spherical wheels. Moreover, the manufacturing process of Mecanum wheels is now quite mature.
The Mecanum wheel resembles a helical gear, with its teeth being rotating drum-shaped rollers whose axes form an angle α with the wheel's axis. This unique structure grants the wheel three degrees of freedom: rotation about its axle, translation along the perpendicular direction of the roller's axis, and rotation about the point of contact between the roller and the ground. Thus, the drive wheel possesses active driving capability in one direction while also exhibiting free movement (passive movement) in another. Instead of a standard tire, the wheel's circumference is composed of numerous small rollers, whose axes are tangent to the wheel's circumference and can rotate freely. When the motor drives the wheel, it moves forward in a direction perpendicular to the drive axle in the usual manner, while the rollers around the wheel rotate freely along their respective axes.
The diagram illustrates the wheel arrangement using an omnidirectional movement mechanism, where the small diagonal lines on the wheels indicate the axial direction of the rollers touching the ground. Each omnidirectional wheel is independently driven by a DC motor. By appropriately combining the rotational speeds of the four omnidirectional wheels, the robot can achieve three-degree-of-freedom omnidirectional movement on a plane. The force analysis of the robot's base, composed of four omnidirectional wheels, is shown in the figure, where σ<sub>0</sub> represents the axial frictional force on the small roller when the wheels are rolling; σ<sub>1</sub> represents the rolling frictional force on the small roller when it is rolling passively; and ω represents the angular velocity of each wheel.
1.4 Design of Robot Leg Structure
The designed leg section consists of two parts: the upper limb and the lower limb. The upper limb connects to the robot's main body and the lower limb, while the lower limb connects to the tires. To enable the robot's legs to meet movement requirements, a joint needs to be designed at the connection between the upper limb and the robot's main body, allowing the leg structure to rotate 360 degrees in the robot's side plane. Furthermore, to control the rotation and its angle, small motors need to be installed at the joint between the upper and lower limbs, so space must be reserved for the installation of the motors and wiring.
As shown in Figure 2, the upper part of the upper limb has a hole and a shaft. The shaft is connected to a motor inside the main body via a coupling, thereby controlling the rotation of the upper limb around the main body. The hole and the vertical hole in the lower part are used to pass electrical wires, and the two holes at the bottom are used to connect to the lower limb. The total length is about 70 cm and the width is about 16 cm.
Figure 2. UG 3D structural diagram of upper and lower limbs
2. Mobile robot motion control system
The motion control system of a mobile robot is the actuator of the robot system, playing a crucial role in the accurate completion of various tasks. Sometimes it can also function as a simple controller. The components of a robot motion control system include: computer hardware and control software, input/output devices, drivers, and sensor systems. Their relationships are shown in Figure 3.
Figure 3 Components of a Robot Control System
The design of a mobile robot motion control system mainly includes the system's functional and architectural design. Functional design primarily involves the software design of control functions and algorithms, while architectural design is the hardware implementation of these functions. The design of the mobile robot motion control system varies depending on the task and environment. Currently, the main problems with robot motion control systems are: the system is limited to dedicated microprocessors and dedicated robot languages, resulting in poor openness; the software structure depends on the microprocessor hardware, making it difficult to port between different systems; and it has poor scalability. To address these shortcomings, the following requirements should be considered when designing a robot motion control system:
(1) Open system architecture. It adopts an open software and hardware architecture, which can be easily expanded as needed to make it suitable for scientific research needs of different purposes;
(2) Reasonable modular design. The hardware is modularly designed according to system requirements and electrical characteristics, which not only facilitates installation and maintenance, but also improves the reliability of the system; the software is divided into different modules according to functions, which facilitates modification and addition;
(3) Real-time and multi-tasking requirements. The controller must be able to process external interrupts within a defined time and can perform multiple tasks simultaneously;
(4) Network communication function, which facilitates resource sharing and multi-robot collaboration;
(5) It has a certain level of intelligence and can make judgments and decisions based on actual conditions, such as handling sudden changes in given speed or when it is outside a reasonable range, and automatically diagnosing faults.
2.1 Robot Drive System
Currently, the most common motion control methods for robots include DC motors, stepper motors, and servo motors. For my project, I need a speed-controllable motor for use as a Mecanum wheel, and another motor with precisely controllable and maintainable angles for use as a leg joint. My initial estimate is that the motor speed is not very high. If I use a DC motor, due to the influence of speed and torque, a reducer will be required, and angle control will be impossible. If I use a stepper motor, a driver will be needed. To meet the system's control requirements, and considering economic factors, I plan to use the Dynamixel AX-12 servo motor, a dedicated servo motor for robots. It can not only precisely control the angle for joint angle control, but also be set to infinite rotation mode via software for use as a wheel.
A servo motor is a position servo actuator suitable for control systems that require continuous angle changes and the ability to maintain that angle. Its working principle is as follows: the control signal enters the signal modulation chip through the receiver's channel to obtain a DC bias voltage. Internally, it has a reference circuit that generates a reference signal with a period of 20ms and a width of 1.5ms. The obtained DC bias voltage is compared with the voltage of a potentiometer to obtain the voltage difference output. Finally, the positive or negative value of the voltage difference is output to the motor driver chip to determine the motor's forward or reverse rotation. The AX-12 servo motor is an intelligent, modular power unit, mainly composed of a microprocessor, a precision DC motor, a gear reducer, a position sensor, a temperature sensor, and a control chip with communication capabilities.
The AX-12 is equipped with an ATmega8 microprocessor, which receives data packets from the controller and processes them before sending PWM signals to the servo motor to control its start and stop. Therefore, controlling the servo motor essentially involves controlling the ATmega8. The servo motor's state and parameters are stored in the corresponding addresses in the ATmega8's RAM and EEPROM; controlling the servo motor is the process of reading and writing data to its corresponding addresses. Table 1 shows the specific parameters of the servo motor.
2.2 Robot's Perception System
Environmental perception is one of the most fundamental capabilities of a mobile robot besides movement, and its level directly determines the robot's intelligence. Perception capability is determined by the perception system, which is a crucial I/O tool for the robot to interact with its environment and humans, serving as the robot's window to information. The reason mobile robots can autonomously move towards targets and complete specific tasks in known or unknown environments is because they can perceive external environmental information and their own state through multiple sensors.
Mobile robot sensors can be divided into two categories: internal and external sensors. Internal sensors detect the robot's own state and are essential for its movement, such as position and velocity sensors; they are indispensable basic components of a robot. External sensors detect the robot's environment and conditions, and their characteristics depend on the task the robot needs to perform. These include vision sensors, ultrasonic sensors, infrared sensors, and sound sensors. The robot uses these sensors to collect various information and then employs appropriate methods to integrate and process the environmental information from multiple sensors to control the robot to perform intelligent operations.
In addition to using the internal sensors such as position, speed, temperature, power supply voltage and torque that are built into the AX-12 servo motor, this design also uses the AX-S1 sensor module as an external sensor.
2.3 Internal Sensors
The AX-12 servo not only has built-in position and speed sensors to detect the motor's rotational speed and the servo's rotational angle, but also internal sensors for temperature, power supply voltage, and torque to monitor the servo's internal condition. When the AX-12 servo's internal temperature, torque, or power supply voltage exceeds the rated range, it actively reports this situation. Furthermore, it will flash LED lights or reduce servo torque to protect itself.
2.4 External Sensors
The Dynamixel AX-12 sensor module is small but complete, containing an infrared distance sensor, an infrared remote control, a sound detection sensor, a light detection sensor, a temperature detection sensor, and even a buzzer.
3. Robot Control System Design
The forces acting on a robot during its movement may change constantly depending on the road conditions, and thus the load on the motor also changes continuously. Therefore, to achieve accurate control of the stroke of the actuator (legs), it is not possible to simply limit the running time of the motor. It is necessary to install a feedback position sensor on the actuator. In this way, when the actuator moves to the specified position, the control system can send a feedback signal to the motor, thereby achieving control of the motor.
Of the six legs, 1, 3, and 5, and 2, 4, and 6 are the same. Therefore, when designing the control, we only need to take the combination of legs 1 and 2 as an example. Legs 3 and 5 are the same as leg 1, and legs 4 and 6 are the same as leg 2.
(1) Sensor of the first leg:
Install limit switches at the extreme positions of vertical swing: upper limit B1, lower limit A1. Install limit switches at the extreme positions of forward and backward swing and at the midpoint: forward limit Z1, backward limit X1, midpoint O1.
(2) Sensor on the second leg:
Install limit switches at the extreme positions of vertical swing: upper limit B2, lower limit A2. Install limit switches at the extreme positions of forward and backward swing and at the midpoint: forward limit Z2, backward limit X2, midpoint O2.
4. Conclusion
For the choice of locomotion method, I first chose Mecanum wheels because their technology is relatively mature and they allow for omnidirectional movement. I then explained their principles. Next was the design of the robot's structure. First, I designed the crucial leg structure, followed by the main body; this process was completed using UG 7.5. After that, I selected the motor type, ultimately choosing servo motors due to their superior functionality. In the design of the control system, I finally created a simple framework without editing the internal instructions.
