Abstract: To realize the functions of face tracking, voice recognition, autonomous movement, action response, and remote control of intelligent home service robots, an embedded device system based on Gene8310 was adopted. The overall design of the home service robot was carried out, expanding the corresponding communication, sensing, and drive systems on the hardware side, and introducing face tracking and voice recognition technologies on the software side, realizing harmonious interaction between the robot and humans. Practice has proven that the system has good real-time performance and robustness. Keywords: Gene8310, home service robot, embedded system 1 Introduction Home service robots represent the future direction of robot development and are a breakthrough point for further strengthening robot technology research in China. Home service robots aim to liberate family members from tedious housework, providing assistance and entertainment. The home service robot introduced in this paper is shown in Figure 1. The robot is 140cm tall and uses a wheeled differential drive for walking. The arms are a three-degree-of-freedom structure simulating the shoulder and elbow joints, using servo motors as joint actuators. The robot is equipped with a camera, microphone interface, touch LCD screen, and ultrasonic sensors. It can interact with family members and answer their questions. The home service robot designed in this paper is based on the Gene8310 embedded device. On this basis, it realizes functions such as autonomous movement, motion performance, face tracking, voice interaction, and remote control. Tests show that the system has good stability, usability, and operability, and can meet the various work requirements of the home service robot. The Gene8310 is an embedded device launched by Intel, which has the characteristics of small size, stable operation, and strong scalability, and supports single 5V power supply. The Gene8310 embedded motherboard is 146mm long and 101.6mm wide, with an integrated Intel Celeron processor with a main frequency of 600MHz, supporting up to 1GB of SODIMM memory slots, and equipped with a CF card slot, hard drive interface, PS2 interface, four USB 2.0 interfaces, two RS232 interfaces, audio input and output interfaces, and can be connected to a CRT/LCD monitor. 2 System Design 2.1 Overall System Design The system design of this paper is shown in Figure 2. [align=center]Figure 2. Design Block Diagram of Robot Control System[/align] Based on the task requirements of the home service robot, the robot control system is divided into the following modules: embedded system, camera, speaker, microphone, face tracking system, voice recognition system, and wireless communication system. The embedded system uses a Gene8310 micro motherboard as the hardware platform and runs the Windows operating system; the face tracking system and voice recognition system are VC-based programs running under the Windows operating system; the wireless communication system is based on the Q2501B GPRS communication module; the motion control is based on the PIC16F877A 7-DOF servo motor control; and the walking control is a differential control drive method for dual walking wheels. 2.2 Motion Control System Design The main motion of the home service robot is to simulate the human arm joints so that the robot can perform some human actions. The robot arm designed in this paper has three degrees of freedom, simulating the human shoulder joint and elbow joint respectively, and can realize actions such as raising the arm, bending the arm, and swinging the arm. The combination of three degrees of freedom of one arm can realize actions such as shaking hands, waving, and beckoning. The combination of two arms can realize basic actions such as clapping and hugging. 2.3 Walking Control System Design The walking control system of the home service robot adopts a dual-wheel differential control drive method. The robot model is shown in Figure 3. Here, we assume that the robot and the ground are purely rolling, and the walking wheels only rotate without slipping, resulting in the kinematic model formula: Where: (x, y) are the reference coordinates of the robot's center point O; θ is the motion direction angle of the robot's center point O; ν is the velocity of the robot's center point O; ω is the angular velocity of the robot's differential turning. Based on the above mathematical model and combined with the robot's structural characteristics, the robot's motion is simplified to the motion of two points in contact with the ground. The positions of the two points determine the robot's position, and the motion state of the two points determines the robot's motion state. In Figure 3, XOY is the global coordinate system, xoy is the robot coordinate system, and O is the instantaneous center of velocity. [align=center]Figure 3 Walking Control of a Home Service Robot[/align] Here, we take the forward direction as the positive direction and the backward direction as the negative direction for unified analysis. Let the speeds of the left and right wheels at a certain moment be vL and vR, and the angular velocities of the left and right wheels be ωL and ωR. Within a very short time interval Δt, the robot's direction and linear velocity can be approximated as unchanged. The distance between the two wheels and the ground contact point (i.e., the span between the robot's two wheels) is l. During walking, the main control system transforms the planned path into the control of the angular velocities of the two independent drive wheels, which change over time. Through a driver and a motor, the two drive wheels are driven respectively, and the angular velocities of the two drive wheels change according to the planned path. When a family member issues a command for the robot to go to a designated location, the robot autonomously reaches the designated destination according to the pre-planned path. 3 Wireless Communication The wireless communication of the home service robot refers to remote communication between humans and the robot. It is best to utilize existing family members' belongings with slight modifications to achieve operation of the service robot. This design utilizes the widely adopted mobile phone SMS to achieve remote communication with the robot. A mobile phone module is added to the robot control system, allowing family members to control the robot via SMS. The GPRS module used is the Wavecom Q2501B, which provides fast, secure, and reliable data communication, voice transmission, and SMS services. It features an Open AT command interface, supports text and PDU-mode SMS, and can be easily operated by sending AT commands via serial port. Notably, the Q2501B module also integrates a GPS module with a positioning accuracy of 3 meters, allowing for convenient transmission of the robot's location information to the terminal via SMS. Remote control of the robot is primarily achieved via mobile phone SMS. Family members outside the home send control commands to the smart home service robot using their mobile phones, and the robot executes the corresponding actions upon receiving the commands, as shown in Figure 4. [align=center] Figure 4 Wireless Communication Module System Block Diagram[/align] 4 Human-Computer Interaction Human-computer interaction with the home service robot includes commands from family members and feedback from the robot to family members. This design focuses on voice interaction between humans and the robot. 4.1 Face Detection This design uses a face skin color model combined with similarity and face contour to detect and locate faces. In the YCbCr color space (where Y represents brightness, Cb represents saturation, and Cr represents hue), the skin color clustering regions in the CbCr subplane differ depending on the Y value, exhibiting a non-linear change in skin color clustering regions with varying Y. Considering the influence of Y values, a non-linear piecewise color transformation is performed on the YCbCr color format, and the transformed image is then projected onto the Cb-Cr two-dimensional subspace to obtain a practical face skin color clustering model. Based on this model, a similarity-based method can be used to detect faces. To eliminate the influence of brightness on face skin color, r = R/(R+G+B) and b = B/(R+G+B) are defined, degrading the three-dimensional primary color space into a two-dimensional half-primary color space. Then, a frequency domain-based homomorphic filtering method is used to perform illumination compensation processing on the image. This significantly improves the accuracy of face detection. 4.2 Speech Recognition The speech recognition engine used in this design is a Chinese semantic recognition system. It can recognize 2000 Chinese characters, which can meet the basic needs of conversation in daily life. Currently, the robot's dialogue with humans can realize functions such as life knowledge, travel knowledge, brain teasers, reciting poems, and chatting, providing entertainment, education, and assistance to family members. 4.3 Motion Response The robot's motion response to humans mainly refers to the robot's response to motion commands issued by humans. In this design, the robot can respond to the following motion commands: forward, backward, left turn, right turn, head shake, please, wave, wave, arm swing, clap, shake hands, hug, goodbye, etc. The process of the robot's action response is shown in the figure below. [align=center] Figure 5 Service Robot Motion Response[/align] When a family member sends a motion request to the service robot through the microphone, the speech recognition engine starts, matches the recognized semantics with the database, obtains the action command code, and then sends the action command code to the control module through the serial port. The robot's actuator then performs the corresponding action. 5. Conclusion This paper designs an intelligent home service robot system based on the Gene8310 micro motherboard. This intelligent home service robot realizes functions such as face tracking, voice recognition, autonomous movement, action response, and remote control. Autonomous movement adopts a dual-wheel differential control drive method, which well meets the requirement of omnidirectional indoor movement. Remote control uses a GPRS network based on Q2501B, which is low-cost, has no distance limitations, and is less affected by the environment, fully utilizing the advantages of wireless networks. However, due to the limitations of existing public wireless network bandwidth, its video transmission is relatively weak. With the development of 3G technology, high bandwidth will enrich the multimedia interactive information of the robot's remote control, allowing family members to remotely view the interior of the home. Face tracking is achieved based on skin color models and facial contour theory, and the model has good real-time performance and robustness. Voice recognition uses a pattern recognition-based voice recognition engine, which achieves good recognition results in specific databases. Future research will incorporate machine learning technology to allow the robot to learn skills in specific home environments. Testing shows that the robot can complete all the basic functions of a pre-defined home service robot. References 1 Jin Xin. Remote monitoring system for mobile robots based on public wireless communication network [J]. Manufacturing Automation, 2006(8):53-58 2 Chen Kaifeng. Research on face detection, tracking and recognition of home service robots [J]. Microcomputer Information, 2006,5-2:228-230