Use of a Multi-Frequency Relay of Ultrasonic Sensors with PIR Sensors to Extend the Sensing Range of an Embedded Surveillance System

Transcription

1 Use of a Multi-Frequency Relay of Ultrasonic Sensors with PIR Sensors to Extend the Sensing Range of an Embedded Surveillance System Ying-Wen Bai, Zi-Li Xie and Chen-Chien Cheng Department of Electrical Engineering, Fu Jen Catholic University, Taipei, Taiwan, 242, R.O.C. bai@ee.fju.edu.tw AbstractIn this paper we design and implement an embedded surveillance system by use of a multi-frequency relay of ultrasonic sensors with multiple pyroelectric infrared sensors (PIR) to extend the sensing range in a home or a storehouse. The ultrasonic sensor module consists of a transmitter and a receiver which are placed in a line direction; however, ultrasonic sensors with the same frequency are disturbed by crosstalk with each other and have a high miss rate. To overcome these disadvantages of the ultrasonic sensor, both ultrasonic sensors and PIR sensors are managed by the majority voting mechanism (MVM). Keywords-Embedded Surveillance System; Majority Voting Mechanism; PIR Sensor; Ultrasonic Sensor through the sensing area. If there is no intruder, the MCU (Micro Controller Unit) will use another ultrasonic frequency to transmit the signal to the next relay station. Use of the relay stations and the frequency conversion extends the sensing range. Our design reduces the environmental interference with the ultrasonic signal. All sensing signals are input to the embedded surveillance system by the GPIO (General purpose input and output), and the MVM program counts the sensing states to determine whether to adopt the MVM or not. If the voting results of PIR and ultrasonic sensor groups pass the criteria, the embedded surveillance system starts the Web camera to capture images. I. INTRODUCTION Surveillance systems have recently become more important for everyones security. The embedded surveillance system, frequently used in a home, an office or a factory [1-3], uses a sensor triggered to turn on a camera [4-5]. Some designs use different types of sensors to achieve reliability by means of the different features of each sensor [6-9]. In this paper we extend our previous design not only by using both multiple PIR sensors and ultrasonic sensors as a sensor group, but also by using the MVM. Ultrasonic receivers and transmitters are located at opposite ends. In our design we use the blocking of ultrasonic transmission to detect whether there is an intruder [10-12]. However, the ultrasonic sensor has a high miss rate from environmental noise in long-distance detection. We use the relay station with multi-frequency conversion in ultrasonic sensors to extend the sensing range while avoiding the cross talk noise [13-16]. II. SYSTEM ARCHITECTURE Fig. 1 shows our design which contains several PIR and ultrasonic sensor groups. The PIR sensor groups obtain the sensing signals from human temperature. In the modules of the ultrasonic sensor groups the transmitter and the receiver are separated. The transmitter circuit generates a multifrequency square waveform, and the receiver circuit amplifies the received signals and s out the noise. When a transmitter transmits signals to a relay station, the ultrasonic receiver determines whether there is an intruder passing Figure 1. Embedded surveillance system with multiple frequency relay. A. Software modules We choose Embedded Linux as our operating system. The program of the majority voting mechanism contains a detection of the GPIO function, a counting and majority voting function, an image captured function and a Web server. The embedded system scans the GPIO sockets, which are connected to external PIR sensors and ultrasonic sensors. To verify the state of each PIR and ultrasonic sensor, the embedded system reads the voltage levels of the GPIO sockets. When the system reads 5V from a GPIO socket, we learn that the PIR sensors or the ultrasonic sensors have been triggered and will execute the majority voting program by counting the state of each PIR and ultrasonic sensor. Majority voting is achieved by the sensor groups of the different GPIO sockets. The embedded system is interrupted by the detection procedure and starts the Web /12/$ IEEE 676

2 camera to capture images. When this is finished, the embedded system starts the detection procedure over again. If the intruder is still in the monitoring area, the count of the GPIO sockets voltage levels continues the majority voting mechanism, and the embedded system again starts the Web camera to capture images. The embedded system uploads the captured images by means of the Web server and the streaming server through the Internet. Figure 3. Ultrasonic receiver circuit. B. Hardware modules We use two groups of the external hardware circuits, the PIR and the ultrasonic sensor group. The PIR sensor produces a weak voltage, so we input the sensed signal to a OP to amplify the weak voltage about 1000 times. The amplified signal vibrates between positive and negative voltage, so we input it to the absolute value circuit, and then we input it to the adjustable comparator to compare the sensing voltage and the reference voltage which are set according to the environment temperature. Fig. 2 shows the ultrasonic transmitter circuit which uses a pulse width modulation (PWM) function in the MCU to send out the desired frequency of the ultrasonic signal. The ultrasonic transducer transforms the voltage waveform into an ultrasonic transmission, and the transducer of the receiver transforms the ultrasonic transmission into the voltage waveform. The PIR sensor senses any changes of the temperature of the environment. We use this sensor in our experiment to detect any change in environment temperature caused by human temperature. The sensing distance of a PIR sensor can reach 4m, and the sensing angle is 110. We place the PIR sensor groups in the ceiling, with the sensing area as a cone-shaped projection. Covering the sensing area with multiple PIR sensor groups extends it, and by using the MVM we enhance the overall sensing probability. But the PIR sensor cannot sense a low-speed or heat-insulated object, because these objects obviously wont change an environment temperature. To compensate for this we use the ultrasonic sensor to sense lowspeed or heat-insulated objects. The ultrasonic sensor has both a transmitter and a receiver which are placed in a line direction for the ultrasonic transmission. When a person passes through the sensing path, the ultrasonic transmission is blocked by the human body. The receiver does not receive the ultrasonic transmission and thus senses a low-speed or heat-insulated object. Figure 2. Ultrasonic transmitter circuit. III. ULTRASONIC INTERFERENCE Ultrasonic transmission is a sound transmission which is easily interfered with by other frequencies. Fig. 4 shows an ultrasonic signal with interference by another frequency. Signal interference causes difficulties in setting up the reference voltage. We have compared the measurements of the transmitter signals from the RC circuit and the PWM circuit and we found that the PWM circuit provides a more accurate and stable ultrasonic signal. Fig. 3 shows the ultrasonic receiver circuit. We use a twostage to enlarge the ultrasonic waveform, and the suppresses any undesired frequency. The MCU comparator determines the level of the ultrasonic waveform, which is the method used for the ultrasonic transmission, whether blocked or not. When there is no intruder, the MCU changes the transmitting frequency of the ultrasonic signal to avoid crosstalk in the next stage. We also design several ultrasonic receivers to receive the ultrasonic transmission and to activate the majority voting mechanism (MVM) to improve the sensing reliability in long-range sensing. The receiving states of all receivers are input to the embedded home surveillance system which uses the MVM depending on the results from the ultrasonic receivers. We measure the transmitter signals between traditional circuit and PWM circuit. The former has a frequency of instability and a high miss rate when adjusting the duty cycle. Figure 4. Ultrasonic signal with interference by another frequency. To investigate ultrasonic interference we set up the experimental environment as shown in Fig. 5 with a different rotation angle. We use other ultrasonic sources with other frequencies to interfere with the main frequency around the receiver. 677

3 Figure 5. Measurement environment of ultrasonic interference with a different rotation angle. We measure the interference by other frequencies at 6 meters. Fig. 6 shows the chart of the measurement results using a 25 khz ultrasonic signal to interfere with other frequencies. the comparator reference voltage, which may be the effect of sound wave diffraction. We therefore measure both the signal and the signal of the blocking and the nonblocking of the transmission at different distances. Tables II-IV show that we fixed 400-times magnification from the s at different distances; the voltage value is changing because the signal is not fixed. Fig. 7 shows the chart of the signal range of the blocking and the non-blocking of the transmission. The red line represents the blocking and the black line the non-blocking of the signal. The range between the blocking and the non-blocking is the range for setting our comparator reference voltage. We have found that the adjustable comparator reference voltage range is very small at a distance of 6 meters and at 7 meters can no longer be determined. To increase the amplitude of the voltage waveform received from the ultrasonic signal of the room we place a PET bottle at the front end for focusing. The PET will focus on the original divergence of the ultrasonic signal to reduce noise interference with the ultrasonic properties of the room wall. Table IV shows our measurements of the amplitude of the voltage waveform of each node in the circuit. TABLE II. BLOCKING OF A 40KHZ TRANSMISSION Figure 6. Chart of the measurement results using a 25 KHz ultrasonic signal to interference with other frequencies. We also compare the interference with 40 khz, 33 khz and 25 khz at 6 meters. The measurements are show in Table I. Twostage after the 6.4V- 8.5V 2.1V-3V 1V-3.5V 0.28V- 0.7V 4m 9V-11V 2.3V- 2.6V 0.8V-3V 0.24V- 0.3V 5m 1.8V-6V 1.4V- 1.9V 1.5V-4V 0.1V- 0.2V 6m 4V-7.5V 0.9V- 1.5V 0.8V-1.7V 0.43V- 0.6V 7m 2V-6V 0.1V-1V 0.5V-2.1V 0.3V- 0.7V TABLE I. COMPARING THE INTERFERENCE WITH 40, 33 AND 25 KHZ TRANSMISSIONS Main frequency Interference 40 khz 33 khz 25 khz frequency 40 khz Strong Weak Weak 33 khz Medium Strong Weak 25 khz Strong Strong Strong IV. IMPLEMENTATION AND EXPERIMENT RESULTS In the experimental results we found according to the specifications of the components that if the ultrasonic sensing distance is more than 6 to 9 meters, and if we give the transmitter the same direction as the sensing direction of the receiver, the ultrasonic transmission is blocked when a person enters the transmission path of the sensing area. The comparator reference voltage is set according to the signal between the blocking and non-blocking of the transmission. An example with a 40kHz transmission at a distance of 6 meters or more shows that as the blocking and non-blocking signal intensity change is very weak, the result is an inability to obtain Figure 7. signals between the blocking and the non-blocking of a 40kHz transmission. 678

4 TABLE III. BLOCKING OF A 33KHZ TRANSMISSION Two-stage 6.0V- 6.24V 4m 5.8V- 6V 5m 5.6V- 5.9V 6m - 5V- 22.8V 5.5V 7m 20V-23V 2.8V- 3.5V 8m 10V-15V 2.8V- 3.5V 10V-19V 2V-17V 7V-13V 6V-10V 5.5V-11V 6V-12V 2V-4V 2V-3V 2V-3V 2V-3V 0.6V-1.8V 2.8V-3.1V Figure 8. signal of the blocking and the non-blocking with focusing mechanism TABLE IV. 4m 5m 6m 7m 8m BLOCKING OF A 25KHZ TRANSMISSION Twostage V V V 5.8V-6V 7V-13V 0.5V-3.5V V 6V-20V 0.5V-3.5V 4V-5.7V 10V-17V 3V-4V 4V~5.7V 17V-20V 3V-~4V 5V-5.6V 13V-20V 3.5V-4V 5.6V- 5.9V 12V-16V 3.2V-4V Fig. 9 shows the measurement of the range of ultrasonic scattering of the environment. Fig. 10 shows the scattering distribution of the plane curves. Our ultrasonic transmitter is placed at 50 cm, the center. If we move from the left to the right, the amplitude of the voltage waveform becomes smaller. An increase in the distance also causes a reduction of the amplitude of the voltage waveform. When the distance is 7 m, even while the ultrasonic signal of the receiver has the same direction as the transmitter, we find that the amplitude of the voltage waveform has been reduced to near the comparator's reference voltage. Hence the scattering causes the amplitude of the voltage waveform to become gradually lower than the reference voltage. To increase the amplitude of the voltage waveform we place a PET bottle at the front end for focusing. Table V shows the physical measurements of the voltage waveform of the at different distances with focusing, and Fig. 8 shows the signal of blocking and non-blocking with focusing mechanism. We have found that the ultrasonic signal increases at the central point and achieves the effect of focusing. We therefore use multiple ultrasonic sensors to receive the ultrasonic transmission to enhance the reliability of this system. TABLE V. BLOCKING WITH FOCUSING 4m 5m 6m 7m Two-stage 9.5V- 12.8V 11.7V-13V 2.6V-7.8V 5.3V-8.8V 4.1V-7V 3.8V- 4.3V 3.3V- 3.7V 1.6V- 2.3V 1.3V- 2.3V 0.6V- 1.7V 2.9V-5.1V 1.6V-4.4V 2.4V-4.8V 1.3V-3.1V 1.2V-2.9V 0.2V-0.3V 0.1V-0.3V 0.3V-0.5V 0.3V-0.7V 0.2V-0.5V Figure 9. The measurement range of ultrasonic scattering of the environment. 679

5 Fig. 12 shows our ultrasonic frequency relay sequence. We also use the PET bottle to focus the ultrasonic signal to reduce the interference by frequencies from previous stages. Figure 10. Curve showing distribution of scattering. Our receiver adds the PET bottle to enhance the focus of the ultrasonic signal. Table VI shows the physical measurements of the amplitude of the voltage waveform, and Fig. 11 shows the scattering distribution of the different deviations of the distance with perpendicular direction. We use multiple ultrasonic sensors to receive the ultrasonic transmission to enhance the reliability of this system. TABLE VI. ULTRASONIC SCATTERING AROUND RECEIVER WITH FOCUSING 4m 5m 6m 7m from the edge 0cm 0.60V 0.53V 0.48V 0.43V 0.39V 10cm 1.26V 0.74V 0.62V 0.56V 0.48V 20cm 1.41V 0.82V 0.76V 0.71V 0.54V 30cm 1.40V 0.96V 0.86V 0.82V 0.56V 40cm 1.81V 1.30V 1.12V 1.01V 0.66V 50cm 2.52V 1.72V 1.41V 1.32V 0.78V 60cm 1.83V 1.31V 1.18V 1.04V 0.64V 70cm 1.44V 1.01V 0.90V 0.83V 0.58V 80cm 1.40V 0.85V 0.81V 0.75V 0.55V 90cm 1.31V 0.78V 0.65V 0.61V 0.51V 100cm 0.81V 0.61V 0.52V 0.46V 0.42V Figure 12. Our ultrasonic frequency switching in the relay sequence. To determine the placement of the relay station we measure the sensing probability at the different distances and frequencies. Our design can adjust the maximal transmitter voltage to 5V and issue 25 khz, 33 khz and 45 khz ultrasonic signals. In the receiver we input the sensed signal to a twostage OP to amplify the weak signal about 400 times, and adjust the reference voltage. Fig. 13 shows that our design can maintain a high sensing probability with a higher sensing range by using the multi-frequency relays. Figure 13. Sensing probability of multi-frequency relays. We use multi-frequencies of the transmitter and the receiver to provide the sensing, and we place the ultrasonic sensors on the walls around the room as shown in Fig. 14. The ultrasonic transmission will be blocked when an intruder enters the transmission path of the sensing area. But the ultrasonic transmission spreads a beam angle which is likely to cause interference, so we use multiple receivers, and the MVM enhances the overall sensing probability. The PIR sensor circuit has three groups of PIR sensors placed on the ceiling. Figure 11. Curves showing distribution of scattering after adding PET bottle focus. 680

6 Figure 14. Arrangement of our experimental environment Table VII shows the overall sensing probability by the relay design and the MVM. In addition we use more relay stations with multiple frequency ultrasonic sensors with PIR sensors to extend the sensing range. TABLE VII. ULTRASONIC SCATTERING AROUND RECEIVER WITH FOCUSING Overall sensing probability of multiple sensor groups Earlier design Our design with relay once and MVM Our design with relay twice and MVM 6 m 65% 100% 100% 12 m 0 % 67% 100% 18 m 0 % 0% 70% V. CONCLUSION Our experiment shows two different types of sensors which are enhancing the overall sensing probability by using the MVM to reduce the shortcomings of both the PIR sensors and the ultrasonic sensors. In addition our design increases the detection range by use of a multi-frequency relay of ultrasonic sensors with multiple PIR sensors to extend the sensing range in a home. REFERENCES [1] Jun Hou, Chengdong Wu, Zhongjia Yuan, Jiyuan Tan, Qiaoqiao Wang and Yun Zhou, Research of Intelligent Home Security Surveillance System Based on ZigBee, International Symposium on Intelligent Information Technology Application Workshops, Shanghai, Dec. 2008, pp [2] Xiangjun Zhu, Shaodong Ying and Le Ling, Multimedia sensor networks design for smart home surveillance, Control and Decision Conference, 2008, Chinese, 2-4 July 2008, pp [3] L. Lo Presti, M. La Cascia, Real-Time Object Detection in Embedded Video Surveillance Systems, Ninth International Workshop on Image Analysis for Multimedia Interactive Services, 7-9 May 2008, pp [4] Wen-Tsuen Chen, Po-Yu Chen, Wei-Shun Lee and Chi-Fu Huang, Design and Implementation of a Real Time Video Surveillance System with Wireless Sensor Networks, VTC Spring IEEE Vehicular Technology Conference, May 2008, pp [5] Mikko Nieminen, Tomi Raty, and Mikko Lindholm, Multi-Sensor Logical Decision Making in the Single Surveillance Point System, Fourth International Conference on Systems, France, 1-6 March 2009, pp [6] Ying-Wen Bai, Zong-Han Li and Zi-Li Xie, Use of Multi-Frequency Ultrasonic Sensors with PIR Sensors to Enhance the Sensing Probability of an Embedded Surveillance System, The 10th International Symposium on Communication and Information Technologies 2010 (ISCIT-2010), Meiji University, Tokyo, Japan October 26-29, 2010, pp [7] Zhuping Qiao, Xiaoli Zhao, Zengxiang Zhang and Qingke Wen. Spatio-Temporal Characteristics Analysis of Land Use and Change in Xinxiang City, International Workshop on Multi-Platform/Multi- Sensor Remote Sensing and Mapping (M2RSM), Jan. 2011, pp [8] Yang Cao, Huijie Zhao, Na Li and Hong Wei Land-Cover Classification by Airborne LIDAR Data Fused with Aerial Optical Images, International Workshop on Multi-Platform/Multi-Sensor Remote Sensing and Mapping (M2RSM), Jan. 2011, pp [9] Lee, J. and Won, C. High Resolution Tactile Imaging Sensor Using Total Internal Reflection and Non-rigid Pattern Matching Algorithm, International Workshop on Multi-Platform/Multi-Sensor Remote Sensing and Mapping (M2RSM), Jan. 2011, pp [10] Hai-Wen Zhao, Hong Yue, and He-Gao Cai, Design of a Distributed Ultrasonic Detecting System Based on Multiprocessor for Autonomous Mobile Robot, Proceedings of the 2007 WSEAS Int. Conference on Circuits, Systems,Signal and Telecommunications, Gold Coast, Australia, January 17-19, 2007, pp [11] Johann Borenstein and Yoram Koren, Error eliminating rapid ultrasonic firing for mobile robot obstacle avoidance, IEEE Transactions on Robotics and Automation, Vol. 11, No. 1, pp , Feb [12] Francesco Alonge, Marco Branciforte, and Francesco Motta, A novel method of distance measurement based on pulse position modulation and synchronization of chaotic signals using ultrasonic radar systems, IEEE Transactions on Instrumentation and Measurement, Feb. 2009, pp [13] Zhong Zhou, Shengli Zhou, Jun-Hong Cui and Shuguang Cui, Energy- Efficient Cooperative Communication Based on Power Control and Selective Single-Relay in Wireless Sensor Networks, IEEE Transactions on Wireless Communications, August 2008, pp [14] Huasheng Zou, Chao Zhou and Gang Lin, Degradation of Chlorobenzene in Novel Configuration Sonoreactors: Comparison Between Sonoreactor with Multi Transducer and Conventional Sonoreactor, ICBBE rd International Conference on Bioinformatics and Biomedical Engineering, June 2009, pp.1-5. [15] A. Novell and A. Bouakaz, Exploitation of multi-frequency excitation for contrast agent imaging, IEEE International on Ultrasonics Symposium (IUS), Sept. 2009, pp [16] T.-L Pham, J.-G Minonzio, M. Talmant and P. Laugier, Impact of a multi-frequency sequence of measurements on first arriving signal velocity on a bone plate model, IEEE International on Ultrasonics Symposium (IUS), Sept. 2009, pp [17] Shan Du, Shehata Mohamed and Badawy Wael. Hard hat detection in video sequences based on face features, motion and color information, International Conference on Computer Research and Development (ICCRD), March 2011, pp [18] Yifan Chen, Wai Lok Woo and Cheng-Xiang Wang. Channel Modeling of Information Transmission Over Cognitive Interrogator-Sensor Networks, IEEE Transactions on Vehicular Technology, Jan. 2011, pp