Research On Data Accuracy Assessment Method In Sensor Networks

As the sensor networks have been becoming larger,the applications and functions of them are becoming more complex in recent years.On the one hand,the accession of a large number of devices has made the network scale larger and sensor data traffic busier.In the meanwhile,the large-scale sensor network has caused the bad storage of sensor data.On the other hand,to meet various demand-responsive decision-making of sensor network terminal applications,the accurate and efficient extraction of sampled data information counts.To establish an efficient connection between the underlying network entity of WSN and the top-level data engineering and decision-making,the first method is to improve the quality of service of the WSN,but it will requir e more on the network hardware resources;the other is to strengthen the processing ability of application layer data and information,but more accurate decision results generally mean more massive data storage and more complex data calculation problems.As a consequence,limitations exist for above two methods.Based on this,this article is aiming to obtain as many high-quality data streams as possible in the case of limited network resources and costs,which will reduce the pressure of data transmission?storage and data calculation.Then,the main work of this paper can be summarized as follows.Firstly,it is important to evaluate the data quality to avoid the generalized collection of WSN data as much as possible,so the evaluation necessity of data accuracy is analyzed.Therefore,a quantitative definition method of data quality based on sensor network data acquisition system has been proposed,and a simplified MEM data processing algorithm has been designed to analyze the pre-sample short time-window data stream.Secondly,the sensor node sampling process has been modeled as an analysis object,and the physical significance of the data quality evaluation for sample has been explained,the original data quality analysis model generated by the data source is established,as well as the relationship between the data quality and the sampling rate.Furthermore,the WSN data packet transmission process has been modeled as an analysis object,and the physical significance of the data quality evaluation for packet transmission has been explained,too.According to the queuing theory,the packet transmission path,the packet loss of the adjacent-node channel error rate is investigated.So,a quality analysis model of the transmission process based on network parameters is established.Meanwhile,an evaluation method of data reliability in a specific area of the network has been given.And taking the flat grid WSN for instance,this paper has designed an example of quality level visualization to provide a global overview of the real-time status of the WSN.Finally,according to the designed data accuracy evaluation method,taking the sample data of the Lora sensor network platform,the above theoretical analysis and design are simulated,and the experimental results can provide a basis for the data quality evaluation of the sensor network.