Vol.3 No.3 2010

Research paper : Development of a sensor system for animal watching to keep human health and food safety (T. Itoh et al.)−229−Synthesiology - English edition Vol.3 No.3 (2010) as mentioned earlier, data reception will be possible without the overhead such as the preamble. Also, by comparing the reception signal intensity of each node, the location of the node at 1 m or less precision can be known. This allows the detection of the outbreak and spread of the disease, as well as the location of the chicken with health abnormality without conducting ID management of the nodes. Another point of shortening the message is the estimation of activity level by reception frequency. For example, in the simplest model, a switch that turns ON when there is an acceleration that surpasses a certain threshold is installed, and one message is transmitted when the number of ONs reaches a certain level[11]. If the activity level is high, the transmission intervals shorten, and when the activity is low, the transmission intervals lengthen. Therefore, the reception interval represents the activity status. Although we cannot tell whether the activity level is abnormal from one transmitted data alone, the activity level monitoring with consideration of the individual differences will be possible by comparing with past data. Using this method, it will not be necessary to include data for the activity level in the message, and ultimately, only the output from the digital thermosensor and parity bit during signal transmission is necessary. For the development of this reception system, the fabrication of the prototype of the simultaneous multi-channel receiver and the confirmation of its operation, the check of the effectiveness of the basic message analysis software, and the development of the algorithm for identifying the node position have been completed. It has been experimentally confirmed that the node location identification can be obtained at 1 m or less precision in an ideal environment.As explained above, downsizing and weight reduction, low power consumption (longer lifespan), and cost reduction are done by simplifying the node through the employment of the high-performance receiver system.5 Experimental infection and prototype systemAs mentioned earlier, one of the important points of this research is to study the disease state of the chickens to optimize the sensor node. At the start of this research project, although it was known that the bird flu that occurred in the Japanese poultry farms could be a threat to humanity, there was no study on how the body temperature changed when the chickens became infected by H5N1. The only data that existed were qualitative data on the daily behavior of chickens in good health and how that may change when infection occurred. The digital MEMS sensor or the event-driven node could be realized only by understanding the properties of the subject, or chicken characteristics in this study, and therefore the research team engaged in experimental infection.In the experimental infection conducted by NIAH, it was shown for the first time from the results of the temperature behavior using the wireless sensor node[14] shown in Fig. 7, that “the fever development and time of death of chickens infected by the highly pathogenic avian influenza virus differ according to the strain”[12], and that “the transmissibility of avian influenza virus among the chickens is correlated to the amount of virus excreted”[13]. The prototype node used in the experiment had exterior size of a one-yen coin and weight of 3 g or less (including weight of the battery) to reduce the burden on the chickens in the experimental infection and poultry house experiment. It also included a thermosensor to measure body temperature and accelerometer to monitor the activity level, as well as the time-driven wireless module that obtained and transmitted the temperature and acceleration data at a certain time interval (can be freely set).In the above experimental infection, the acceleration data as well as the temperature data were recorded. Based on these finding and data, an infection determination program was developed using both the temperature and activity level patterns. This was compared to the conventional data, and it became possible to discover (recognize) the possibility of infection early and automatically. Moreover, the number of probes and individual thresholds can be set in the digital thermosensor, and the activity level sensor can be set with appropriate acceleration threshold. Also, the virus Fig. 6 Comparison of the ULP wireless sensor and the conventional sensor New design (ULP) typeConventional type③Not necessary④Achieving low power consumption of node (degree of importance 20 %)②Achieving all-in-one and customization (degree of importance 25 %)①Achieving digital MEMS (degree of importance 30 %)⑥Achieving downsizing and flexibility(degree of importance 10 %)⑤Achieving flexibility (degree of importance 15 %)AntennaCustom RF-ICActivity levelsensorThermosensorBatteryPrinted circuit boardRF-ICAccelerometerThermosensorAmplificationcircuit etc.MicrocontrollerPrinted circuit boardSensor interface circuitSimple processing (logic) circuitHigh frequency transmission circuit<10 nW0 nW<1 µWBatteryAntenna


page 44