Performance Analysis Based On Kalman Filter With Different Quantitative Strategies

Wireless sensor network (WSN) has many advantages such as low cost, flexible operation, highreliability and so on. However, it is vulnerable to the bandwidth constraint in the transmission channel.Therefore, it is the key problem for sensor network research to reduce the communication bandwidthrequirements（reduce the requirements for communication bandwidth） in guarantee quantization error ofmeasurement data in a range of conditions or effectively reduce the quantization error of the measurementinformation in a certain bandwidth constraints.Due to the physical characteristics of the sensor network itself, information in the process of networktransmission is often need to carry out quantitative processing. Researchers have proposed some effectivequantitative filter(s), but people often face choices such as quantitative objects and quantitative rateproblem in the process of application.In view of the above, this letter makes some studies for quantitative filter in terms of quantization error,quantitative copies, and quantitative range. The main research contributions and innovation points are asfollows:Firstly, reasonable quantitative range is calculated under the condition of same quantitative copies. Inthis letter, the graphs of quantitative objects are drawn in terms of changes of quantitative objects data, andthen the quantitative upper and lower bounds are studied based on quantitative objects data and graphs.Under the condition of same quantitative copies, this letter concludes the changing rule of the quantizationerror in terms of quantitative range widening and narrowing and chooses the quantitative interval with thesmallest quantitative error as the reasonable quantitative interval.Secondly, reasonable quantitative filter is chosen under the condition of certain quantization error. Inthis letter, two quantitative filters’ changing rule of the quantization error is compared. Quantitative copiesthat are required by two quantitative filters to meet quantitative errors respectively can be computed byknown quantitative errors. Hence, it provides evidence for the selection of filter.Thirdly, compute the suggested quantitative copies under the condition of certain bandwidthlimitations. In this letter, the maximum quantitative copies of the transmission network in this system arecomputed. Under the condition of same quantitative interval, the quantitative copy increases from one to the maximum in order to compute the quantization error and draw its graph. Through the analysis, there is aturning point in the process of quantification number increasing; the decreasing rate of quantization errorgets smaller significantly, when the quantification number is greater than the turning point. This turningpoint is the suggested quantitative copies.Finally, in many dimensions system, reasonable quantitative range of each dimension has beencomputed, this letter makes a comparative analysis of each dimension’ estimation error change in terms ofquantization error, quantitative copies, quantitative range under two conditions of same quantitative rangeand different quantitative range.