Interactive Assistive Electric Wheelchair with Embedded System

• Main Authors: Yin-ChenWang, 王尹辰
• Other Authors: Ching-Hsing Luo
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/91541819120501467897

碩士 === 國立成功大學 === 電機工程學系 === 102 === 1.SUMMARY Based on the technology of high-efficiency dual-core embedded systems, the authors propose an integrated electric-powered wheelchair system that integrates related hardware peripherals and sensors such as Wi-Fi adapters, electric wheelchair motors, linear actuators and wireless video transmissions. To transmit video immediately and effectively, a JPEG image is compressed and encoded on a dual-core embedded system platform using a codec engine software architecture approach. The hardware configuration technology uses shared memory to transfer ARM and DSP data and employs fixed-point calculation on the fixed-point DSP to increase its efficiency of calculation. The advantages of dual-core embedded system platform features and design multithreading achieve the effect of parallel processing. In terms of safety for electric wheelchairs, the system uses the proportional-integral-derivative (PID) control strategy to improve deficiencies in going straight because the dynamic characteristics of left and right wheel speed are different. Furthermore, real-time image transfer allows the images from a wheelchair’s blind side to be transferred to the user’s tablet PC, thus increasing safety. 2.INTRODUCTION In the face of an aging society, many people are experiencing age-caused diseases and gradually losing the ability to move, which means they must rely on electric-powered wheelchairs for independent movement. Currently, the most common type of input device for electric-powered wheelchairs is the joystick, yet people whose hands cannot operate normally are unable to use such a device. Therefore, current research is aimed at finding better input devices for electric-powered wheelchairs. Action identification achieved through voice control [4] and head movement with the use of a webcam [9] are some of the ideas that have been discussed. But these input devices are often designed for a specific disease and are not available for common use among people with all kinds of severe disabilities. Furthermore, the hardware design of these new devices use fixed sets of input, so if there is ever a need to replace other input methods, a lot of effort and time must be spent. In the past, the development of electric-powered wheelchairs mostly focussed on the mechanism—such as the folded vs. the reclining wheelchair—and the integration of the joystick for control was taken for granted. The joystick device is mostly presented as a part of the hardware rather than, weakly, the software. Yet, if an electric-powered wheelchair is assisted by software, integrated with the necessary hardware and updated with an improved mechanism and technique, future electric-powered wheelchairs—whether using external controls or auxiliary interactive features—will match the demands of many different physically challenged groups. With vigorous technology developments, a large number of embedded systems have recently been used in mobile phones, digital cameras, network equipment and even high-level home appliances; smart phones and tablet PCs are becoming more common—almost everyone has one. If we use these devices as a user interface and use Wi-Fi to transmit relevant signals to a microcomputer and the drive circuits associated with electric-powered wheelchairs, then tablet PCs can communicate directly with electric-powered wheelchairs. In this way, tablet PCs may be applied to assist users with control operation and interaction. Moreover, regarding version control, software and firmware can easily be updated in this way. However, people who have lost mental and physical abilities cannot control tablet PCs without specific help. So, we use the Morse code converter [1], [2], [5] (which was developed by our laboratory) as an input device. This device simulates the mouse and keyboard for tablet PCs. People with disabilities can take advantage of various different input switches—for example, mouth control, foot control and head control—and can pass the signal through the Morse code converter to the tablet PC application to manipulate electric-powered wheelchairs. With the assistance of a high-performance micro-controller, we integrate the control circuit with real-time video transmission in addition to other necessary control items (electric wheelchair motor, linear actuator or optical encoder). This transmission will have direct communication with the user’s tablet PC or mobile phone and enable interaction with the user to increase driving safety, thus meeting user’s needs and providing a smart electric-powered wheelchair for users with severe disabilities. 3.MATERIALS AND METHODS Embedded auxiliary within the proposed interactive system for electric wheelchairs allows people with severe physical or mental disabilities to control their electric-powered wheelchairs and is also interactive and multi-functional. Morse code is input through a mouse-controlled switch via a Morse code converter, which converts various signals into signals for mouse or keyboards, thus operating applications in tablet PCs. Then, through PC applications, the electric wheelchair is controlled using Wi-Fi technology and electronic control systems that allow electric wheelchairs to communicate with each other. Among the electronic control systems in this electric-powered wheelchair, two microprocessors are used for system integration—namely, BeagleBoard-xM and STM32F407. The main microprocessor is BeagleBoard-xM, which serves as the main core of the entire electronic control system and is in charge of every peripheral’s state information. Communication between the electric-powered wheelchair and the user’s tablet PC goes through BeagleBoard-xM first; therefore, we use a USB Wi-Fi adapter for two-way communication with the user’s tablet PC. In addition, to allow users to watch ground images from the electric-powered wheelchair, a webcam is added to BeagleBoard-xM to transmit images to the user’s tablet PC in real time. Regarding image transmission, codec engine technology is used to call the BeagleBoard-xM DSP core to take complex calculations to the DSP core and to reach communications between DSP and ARM. The secondary microprocessor is STM32F407, which is developed and designed specially for control; therefore, it plays the role of directly communicating with peripheral control such as motor controllers for the electric-powered wheelchair, linear actuators and rotary encoders. BeagleBoard-xM and STM32F407 are in a master–slave relationship using UART signals for serial transmission. The mobile app we developed allows a graphic interface to use the commercial tablet computer or smartphone as a user control panel using Wi-Fi transmission with SoftAP technology to realise communication with electric-powered wheelchair systems. Moving information and conditions can provide feedback to the user in real time, thus providing guidance assistance and interaction. Graphics interface is available on the tablet computer or smartphone in the market as the user control panel with the mobile app we developed using Wi-Fi transmission with SoftAP technology to realise communication with electric power wheelchair systems. Moving information and condition can provide feedback to the user in real-time and reach its assistance and interaction effects in the way of guiding. 4.RESULTS AND DISCUSSION Experimental results show that this system can effectively transmit images from electric-powered wheelchair systems to tablet PCs or smartphones for display using Wi-Fi wireless transmission. For instance, taking a resolution of 320 × 240 as an example, the executive speed of the electric-powered wheelchair system can handle approximately 29.4 frames per second, while the receiving end of the user’s tablet PC can receive 27.6 frames per second. The control system of an electric-powered wheelchair uses a PID controller to adjust appropriate parameters and obtain stability; this is achieved after two rounds. In one experiment, subjects with body weights of 74 and 81 kg, respectively, used the electric-powered wheelchair system to move a distance of two meters; in both cases, the system had to adjust suitable PID parameters and remain steady and stable without any offset in linear motion. In the case of the 74 kg subject, the average error for two wheels was 0.65 cm; in the case of the 81 kg subject, it was 0.41 cm. 5.CONCLUSION This study combines high-order embedded systems with electric-powered wheelchairs. Considering the popularity of smart phones and tablet PCs, this study proposes a system architecture for smart electric-powered wheelchairs. Wheelchair users benefit from a good user interface and interaction, and caregivers may also use their own smartphones or tablet PCs to view their patient’s operational situation and images in real time. Two microprocessors (i.e. BeagleBoard-xM and STM32F407) are used in the electric-powered wheelchair’s hardware to constitute the entire electrical control system. The more powerful functions and complex work (image transmission, Wi-Fi SoftAP, etc.) are in the charge of BeagleBoard-xM, while STM32F407 is responsible for the control systems of peripherals (wheelchair motor drives, rotary encoders, PID feedback, etc.). With the electric-powered wheelchair user control interface, participants can use their smartphone/tablet system and make use of Wi-Fi transmission to control the wheelchair. The interface application is Android-based, designed in a lightweight software to achieve the best operating effect and to reduce the burden of application resources.