POSITION CONTROL OF VISUAL PROSTHETIC SYSTEM USING ARDUINO PROCESSOR

Transcription

1 International Journal of Mechanical Engineering and Technology (IJMET) Volume 9, Issue 9, September 2018, pp , Article ID: IJMET_09_09_124 Available online at ISSN Print: and ISSN Online: IAEME Publication Scopus Indexed POSITION CONTROL OF VISUAL PROSTHETIC SYSTEM USING ARDUINO PROCESSOR Anis Kazi, Sourabh Arbune, N. Ruban and V. Bagyaveereswaran School of Electrical Engineering, VIT, Vellore, Tamil Nadu, India ABSTRACT Detection and recognition of the individual standing in front of a visually challenged person completely depends on the response given by the person standing. The independent person identification system will help the disabled people in large extent. The main objective of the proposed work is to control the position of visual prosthetic device and to detect the individual standing in front of the visually challenged person. Here, Arduino processor is used for position control of the prosthetic device. For face detection purpose, this proposed work uses HAAR algorithm which is written in C# language. The inputs for the different faces can be taken from the camera module which is fixed with a motor which will be driven with the help of the processor. The prosthetic device is designed in such a way that it can detect person talking in front of challenged person. In real time the device is capable of detecting individuals by comparing database given and effectively recognizes the identity of the individual. The main aim is to assist the challenged person to overcome the dependency, so that the individual can feel free to navigate in outside world. Keywords: HAAR Algorithm, Visual Prosthesis, Face Recognition, Position control. Cite this Article: Anis Kazi, Sourabh Arbune, N. Ruban and V. Bagyaveereswaran, Position Control of Visual Prosthetic System Using Arduino Processor, International Journal of Mechanical Engineering and Technology, 9(9), 2018, pp INTRODUCTION People with complete blindness or with low vision often have difficulty navigating outdoors or in crowded area which increases their dependency on others. They often bring their family or friends with them so that they can help them navigate freely in outside world. Nowadays, cities are growing with the speed of light with increasing number of vehicles, city transportation along with crowd. Visually challenged people finding it very difficult to adapt in such conditions and this also increases their dependency. To overcome this problem visual prosthetic device is developed. This paper studies the control of position of Visual Prosthetic Device by using Arduino Processor. Arduino Processor is given a pre-defined command which it acts accordingly. For this project the database is created in C# platform. Database contains various human facial editor@iaeme.com

2 Position Control of Visual Prosthetic System Using Arduino Processor pictures which is used to compare it with the captured images taken by the camera. Prototype consist of a camera for capturing images, a ultrasonic sensor for sensing the objects, a servo motor for rotation of camera and a MIC for sensing the voice. The device is mount on visually challenged person s body. As the device is compact and light weight it is convenient for the person to handle it. Figure 1 Visual Prosthetic System When the voice given by unknown person is sensed by mic, input will be given to the processor. Processor will command the servo motor to operate and rotate according to the direction of voice received. Ultrasonic Sensor which is mount on motor will simultaneously start sensing the objects or person. As the person is sensed by the ultrasonic sensor, the controller will command the camera to click pictures accordingly. This pictures clicked by the camera will create a new data base within it and starts comparing it with the default database given by the programmer. This proposed work uses HAAR algorithm for Face detection and Reorganization. HAAR algorithm divides the picture taken by camera and divides it into several parts and finds the human facial parameters like nose, mouth, eyes and declare it a human picture or not. This picture will compare these parameters with the default database parameters and display the particular name given for that picture. This proposed system will give a result through earphones or speakers to the visually challenged person. This will let the disable person know or recognize the unknown person. If the person s parameters do not fulfil the requirements needed, the processor will declare it as unknown person. This paper is organized as follows: firstly different parameters of all components related to hardware are discussed. Secondly face recognition software including HAAR algorithm is elaborated along with the portable device programming for Arduino. Finally the approach for controlling the visual prosthetic device with different outcomes is presented. 2. METHODOLOGY 2.1. Hardware Considering the technical view, various components have different ratings and specifications. Fig.1 shows all the components used in this work. An Arduino UNO development board [1] is used with 14 numbers of digital inputs and outputs, of which 6 can be used as PWM output. There are 6 analog inputs in microcontroller board along with 16MHz quartz crystal, USB connection, reset button, power jack and In Circuit Serial Programming header. The board comes with RAM size of 2048 bytes and program memory size of 31.5 kb [14]. Minimum operating voltage is 5 volts. The Arduino UNO was chosen because of its ease of operation, compact size ( ), light weight (28g), open source framework and low cost. A VGA camera module with interface auto exposure active for Arduino is used. It is having lens size of 1/6 inch and vision angle of 30 degree s [10] and [11]. It is chosen because of its editor@iaeme.com

3 Anis Kazi, Sourabh Arbune, N. Ruban and V. Bagyaveereswaran compatibility with hardware, driver availability and video streaming ability at pixel resolution. A servo motor is used with operating voltage of 4.8 to 6 volts dc and average speed of 60 degrees in 0.20 seconds (at 4.8 volts)[1]. It has a torque of 8.5 kg/cm at 4.8 volts. Camera module is installed on motor. For object detection purpose an ultrasonic sensor is used with a range of 2cm to 450cm with an angle less than 15 degrees [12] and [13]. Figure 2 Schematic of the proposed system Figure 3 Front view of Prosthetic Visual Device Fig.3, Fig.4 and Fig.5 show the visual prosthetic device along with the Microcontroller development board. It shows the camera mounted on servo motor which rotates in 180 degrees. A USB cable is available for integration of camera module and development desktop PC. The figures shown explain the hardware circuitry of the proposed methodology in detail. All the components and their position can be clearly observed from the hardware pictures shown in all the figures. The side view, front view and the top view is explained in the given figures. The microcontroller development board is connected to the development desktop PC through an USB cable. Once it is programmed the user can carry the device easily without any complications [2]. The future development in program can be done by just connecting the device and development desktop PC and it doesn t need any external aid editor@iaeme.com

4 Position Control of Visual Prosthetic System Using Arduino Processor Figure 4 Side view of Visual Prosthesis Device The microcontroller development board and servo motor along with camera module is connected with lead wires. Ultrasonic sensor is mounted on camera module. Figure 5 Top view of Visual Prosthetic Device 2.2. Face Recognition Software The operating system chosen for face detection was windows. It uses window visual studio software for compiling C# codes. The camera module is integrated with development desktop PC. The sampling program used was written in C# language which uses HAAR algorithm for face recognition [8] and [9]. The block diagram of HAAR algorithm used for face recognition is given in Fig. 6. The main emphasis was given to coding program into sequence of logical steps. After starting up the device, the camera would start searching for the person or object editor@iaeme.com

5 Anis Kazi, Sourabh Arbune, N. Ruban and V. Bagyaveereswaran Figure 6 Face Recognition using HAAR A predefined database was available on development desktop PC. After detecting the person sampling program will compare the image captured with the database. It uses HAAR algorithm for face recognition. For supporting HAAR Algorithm, it uses Open CV and emgu CV [7] and [10]. These both are libraries containing the codes that support HAAR Algorithm. Open CV is library containing programming functions particularly utilized for real time computer vision. Emgu CV is cross platform.net wrapper which is used for open CV processing library. C#, VC++, Python, etc. are compatible languages for.net which will be used by open CV [8] and [9]. The common features for the face detection and the procedure is given in Fig. 7(a) and (b). The algorithm for HAAR face recognition system is: After detecting the person it captures an image. It divides the image in four equal parts and again starts searching for human facial parameters. Once the parameters get detected, it compares it with database. If parameters matches then it will display the name otherwise it will display stranger. (a) (b) Figure 7 Common HAAR Features and summed area of Integral Images editor@iaeme.com

6 Position Control of Visual Prosthetic System Using Arduino Processor 3. PORTABLE DEVICE PROGRAMMING Arduino Programming can be developed in C#, Matlab, or C programming [5]. For this prototype C platform is used. When prototype starts operating, Ultrasonic Sensor which is mounted on motor will simultaneously start sensing the objects or person. Arduino coding is written in such a way that it will command the motor to rotate until the object or person gets detected [14]. Figure 8 Flowchart for face recognition The flowchart shown in Fig. 8 gives the brief explanation of the compilation steps of the portable device program. The process starts with real time image capturing followed by the comparing the captured images with given sets of preloaded images. If the captured image and preloaded images matches then it displays the name of the person whose image matched with the dataset using HAAR algorithm. If the image doesn t match with the dataset then it displays the person name as stranger. The displayed name is then converted into a voice signal and given to the speaker available in Development desktop PC. 4. RESULTS In development, a desktop based database is created from which the HAAR Algorithm [6] and [7] is taking references. In the database, facial parameters of different individuals are coded from which facial recognition program is able to recognize the individual. The actual output of the proposed system is given in Fig. 9. In this paper, prototype is using a camera which is streaming a video of 640*420 pixel density. It captures an image which is processed through HAAR Algorithm [4] and [7]. It detects the face on individual and compares it with database and gives the command accordingly. If the facial parameters of individual have not been added in database, the program will display the output as stranger after detecting his/her face. Low rating motor was suitable which can rotate the camera through 180 degrees [2]. It also makes the device light weight which leads to ease of handling of the device. Using higher rating motor will give more precise and faster response [3]. The motor is expected to rotate in particular direction which is either right or left until it detects an object or person. Following is the expected observation showing motor rotation in either direction with respect to time editor@iaeme.com

7 Anis Kazi, Sourabh Arbune, N. Ruban and V. Bagyaveereswaran Figure 9 Output of Face Recognition program Table 1 Expected outcome of angular rotation of servo motor with respect to time SR.NO Time in sec Angle towards right Angle towards in degrees left in degrees DISCUSSION AND CONCLUSION In this paper, authors tested the possibility by using commercial availability of Arduino along with camera module and servo motor compatibility with low cost, compact size and light weight. By using algorithms based on face recognition, a modified camera which implanted on motor and optimizing hardware using less power for its operation. It is possible to make a lightweight, flexible, portable device capable of capturing images from real time video streaming. Various techniques have been discovered in last few years for image processing. The authors have used HAAR algorithm in this proposed work. The advantage of using HAAR like feature is to reduce the calculation time. As it divides an image into parts it has faster response time. It uses open CV and emgu CV libraries which support Table 1: Expected outcome of angular rotation of servo motor with respect to time ts image processing algorithm. Open CV is library which supports HAAR algorithm. Emgu CV is cross platform.net wrapper which is used for open CV processing library. C#, VC++, Python, etc. are compatible languages for.net which is used by open CV. Different microcontroller development boards can be used instead of Arduino board like raspberry pi, ARM processor, etc. which will give different parameters. It will also vary the cost of overall system. With this rapid change in performance future devices will have better performance and higher resolution image acquisition. Arduino development board is inexpensive in comparison with other microcontroller boards like raspberry pi and ARM processor. Arduino software is also compatible with different operating systems like windows, Macintosh OSX and Linux. Programming environment for Arduino is easy for beginners as well as for high end users. Arduino software is an open source platform and can be extended by different C++ libraries. Servo motor was used because of its precise control of editor@iaeme.com

8 Position Control of Visual Prosthetic System Using Arduino Processor angular moments. It has close loop control with high torque, high speed and good efficiency [15]. REFERENCES [1] Fink, W., You, C. X., & Tarbell, M. A. (2010). Microcomputer-based artificial vision support system for real-time image processing for camera-driven visual prostheses. Journal of biomedical optics, 15(1), [2] Vishnu, V. S., Aneesa, K. A., Lal, A., & Nabi, A. (2016). Real Time DC Motor Speed Control Using PID in LabVIEW with Arduino. Imperial Journal of Interdisciplinary Research, 2(5). [3] Khan Masoom Raza, Mohd. Kamil, Pushpendra Kumar (2016). Speed Control of DC Motor by using PWM. International Journal of Advanced Research in Computer and Communication Engineering, 5(4), pp.no [4] N.Bagyalakshmi, R.Abinaya (2016). Speed and Direction Control of DC Motor Without using Microcontroller. International Journal of Advanced Research in Electrical, Electronics and Instrumentation Engineering, 5(1), pp.no.1-5. [5] Prithviraj R. Shetti, Ashok G. Mangave (2014). DC Motor Speed Control with Feedback Monitor based on C# Application. International Journal of Research in Engineering and Technology, 3(3), pp.no [6] Vikhe, P., Punjabi, N., & Kadu, C. (2014). Real Time DC Motor Speed Control using PID Controller in LabVIEW. International Journal of Advanced Research in Electrical, Electronics and Instrumentation Engineering, 3(9), [7] Bourbakis, N. (2008, November). Automatic Image-to-Text-to-Voice Conversion for Interactively Locating Objects in Home Environments. In Tools with Artificial Intelligence, ICTAI'08. 20th IEEE International Conference on (Vol. 2, pp ). IEEE. [8] Guo, H., Qin, R., Qiu, Y., Zhu, Y., & Tong, S. (2010). Configuration Based Processing of Phosphene Pattern Recognition for Simulated Prosthetic Vision. Artificial organs, 34(4), [9] Tsai, D., Morley, J. W., Suaning, G. J., & Lovell, N. H. (2009). A wearable real-time image processor for a vision prosthesis. Computer methods and programs in biomedicine, 95(3), [10] Schapire, R. E., & Singer, Y. (1999). Improved boosting algorithms using confidencerated predictions. Machine learning, 37(3), [11] Ma, S., & Bai, L. (2016, August). A face detection algorithm based on Adaboost and new Haar-Like feature. In Software Engineering and Service Science (ICSESS), th IEEE International Conference on (pp ). IEEE. [12] Olden, P., Robinson, K., Tanner, K., Wilson, R., & Basher, A. H. (2001). Open-loop motor speed control with LabVIEW. In SoutheastCon Proceedings. IEEE (pp ). IEEE. [13] Matteucci, P. B., Byrnes-Preston, P., Chen, S. C., Lovell, N. H., & Suaning, G. J. (2011, August). ARM-based visual processing system for prosthetic vision. In Engineering in Medicine and Biology Society, EMBC, 2011 Annual International Conference of the IEEE (pp ). IEEE. [14] Cristinacce, D., & Cootes, T. F. (2003, September). Facial feature detection using AdaBoost with shape constraints. In BMVC (pp. 1-10). [15] Menezes, P., Barreto, J. C., & Dias, J. (2004). Face tracking based on haar-like features and eigenfaces. IFAC Proceedings Volumes, 37(8), editor@iaeme.com

9 Anis Kazi, Sourabh Arbune, N. Ruban and V. Bagyaveereswaran [16] Archana H. Sable and Dr. Girish V. Chowdhary, A Two Phase Algorithm For Face Recognition In Frequency Domain, International Journal of Computer Engineering and Technology (IJCET), Volume 4, Issue 6, November - December (2013), pp [17] U.K. Jaliya and J.M. Rathod, A Survey On Human Face Recognition Invariant To Illumination, International Journal of Computer Engineering and Technology (IJCET), Volume 4, Issue 2, March April (2013), pp [18] Ram Ratan Ahirwal, Sagar Nachankar and Y. K. Jain, A Novel Method Of Average Filtering For Removing Noise And Face Recognition, International Journal of Computer Engineering and Technology (IJCET), Volume 5, Issue 8, August (2014), pp [19] Walaa M Abdel-Hafiez, Mohamed Heshmat, Moheb Girgis and Seham Elaw, A New Face Recognition Scheme For Faces With Expressions, Glasses And Rotation, International Journal of Computer Engineering and Technology (IJCET), Volume 5, Issue 4, April (2014), pp editor@iaeme.com