INFLATE: Incremental wireless transmission for sensor information in industrial environments

Abstract

Abstract-As smart factory trends gain momentum, there is a growing need for robust information transmission protocols that make available sensor information gathered by individual machines and enable their algorithmic exploitation. Wireless transmission enables often-called-for flexibility, yet it poses challenges for reliable and timely transmission of information. This paper proposes a wireless transmission scheme for sensor information of production cycles in industrial environments. We include a preview functionality based on a discrete cosine transform that allows for rapid detection of problematic characteristics. The transmitted information's precision is improved using incremental updates as wireless capacity permits. Further, we include compact meta data that allows receivers to bound the received information's error. Evaluation results show that, even with high packet loss, characteristic features of sensor information are available early, and that error bounds closely follow the actual error.

Effective Lossless Compression of Sensor Information in Manufacturing Industry

“…In our use case, power is readily available at a machine; wireless channel capacity is still limited, though. Moreover, smart factories measure a wide variety of different physical quantities, and the recordings of different machines' sensors are less correlated [22].…”

“…By separating the signal's AWGN noise in this way, the resulting AC model achieves high precision. Our selection mechanism improves over other strategies, which transmit only low-frequency coefficients [22], select coefficients based on relative energy [21], or quantize high-frequency components with decreasing precision [35]. Formally, we define a selection function f k that maps coefficient vectors from D n to an index set of size k such that the sum of the selected coefficients' energy is maximal for a fixed k.…”

Best of Both Worlds: Prioritizing Network Coding without Increased Space Complexity

“…In previous work [7], we have demonstrated how DCT can be applied to sensor information to divide it into different layers with incremental quality. Similarly, for most forms of video or audio streaming applications, a suitable prioritization already exists in the form of a base layer and enhancement layers.…”

“…For some video streaming applications, few layers -e. g. a base layer and a small number of layers providing video quality increments -may provide sufficient flexibility. Other applications, such as the protocol discussed in [7], benefit from a large number of layers to provide many subsequent increments of information quality.…”

“…In addition to video streaming, HNC can be applied to all applications that can prioritize information and benefit from early reception of more highly prioritized information. For example, information that is sampled by remote sensors can be transformed using Discrete Cosine Transform (DCT) to obtain similar enhancement layers [7].…”