Road surface wetness is a contributing factor in traffic accidents. As the amount of friction reduction correlates with the water film height covering the road surface, a quantification is of high relevance in order to improve traffic safety. Both drivers and autonomous vehicles would benefit from additional information. This paper presents a novel concept for road wetness quantification. It is based on a $2\times 4$ -planar capacitive transducer array, capable to detect water spray ejected by the tires and its wetness-related dependencies. The reliable assessment of these dependencies by a proposed capacitive sensor system is shown in an experimental study on an asphalt circuit for various wheel speeds. Besides the spray’s correlation with speed, the results reveal significant differences in transducer positions and designs confirming the array’s relevance regarding wetness quantification. In addition, a 1-nearest neighbor classifier capable of automatically distinguishing between eight wetness levels is proposed. The classifier is optimized by an extended version of balanced accuracy and reaches similar performance as binary classifiers from related research. A balanced ratio between capacitance increase-, standard deviation- and speed-related feature types is one key aspect of classifier performance. Furthermore, up to a certain extent, the array’s individual transducers can significantly contribute to classifier performance with design- and position-related advantages.

The use of artificial intelligence (AI) in sensor analytics is entering a new era based on the use of ubiquitous embedded connected devices. This transformation requires the adoption of design techniques that reconcile accurate results with sustainable system architectures. As such, improving the efficiency of AI hardware engines as well as backward compatibility must be considered. In this paper, we present the Hybrid-Float6 (HF6) quantization and its dedicated hardware design. We propose an optimized multiply-accumulate (MAC) hardware by reducing the mantissa multiplication to a multiplexor-adder operation. We exploit the intrinsic error tolerance of neural networks to further reduce the hardware design with approximation. To preserve model accuracy, we present a quantization-aware training (QAT) method, which in some cases improves accuracy. We demonstrate this concept in 2D convolution layers. We present a lightweight tensor processor (TP) implementing a pipelined vector dot-product. For compatibility and portability, the 6-bit floating-point (FP) is wrapped in the standard FP format, which is automatically extracted by the proposed hardware. The hardware/software architecture is compatible with TensorFlow (TF) Lite. We evaluate the applicability of our approach with a CNN-regression model for anomaly localization in a structural health monitoring (SHM) application based on acoustic emission (AE). The embedded hardware/software framework is demonstrated on XC7Z007S as the smallest Zynq-7000 SoC. The proposed implementation achieves a peak power efficiency and run-time acceleration of 5.7 GFLOPS/s/W and $48.3\times $ , respectively.

Due to the correlation between friction reduction and the road-covering water film height, knowledge about the current wetness level is of relevance for drivers and autonomous systems. A promising approach for wetness quantification is based on capacitive transducer arrays that enable to detect water spray ejected by the tires. Even though previous studies on this approach have shown the feasibility of wetness classification, optimization opportunities exist. While these previous investigations were limited to one feature selection algorithm, we study various algorithms and demonstrate the potential for optimization. Besides an application-specific algorithm that is capable of determining class-dependent features resulting in a performance improvement of more than 0.03 for the considered evaluation metrics, sequential forward floating selection in particular yields the most significant performance increase of more than 0.06. In addition, prior studies were limited to a test track with constrained conditions. Thus, in order to study the transferability of the preceding results, we present investigations on new experimental data acquired on public roads. The unknown and varying environmental conditions on public roads as well as a larger wheel speed range and steering angle effects are shown to significantly decrease classifier performance. We demonstrate that fusing two transducer arrays of different wheel arch liners increases performance by around 0.02. Here, a considerable information benefit can be attributed to the different transducer positions with design-related advantages. Furthermore, we show that the fusion with additional sensor data available in the vehicle results in a further performance improvement of more than 0.02.

The present work investigates the estimation of rotational speeds based on neural networks and high sampled gearbox vibration data. A total of 18 different network architectures consisting of fully connected network layers, long short-term memory cells and different activation functions are compared. In addition, the impact of the sampling rate of the vibration data, the signal length and the input data type is investigated. Vibration data in the time domain, frequency domain and a combination of both are examined as input data. Finally, the number of neurons is varied to obtain the best possible rotational speed estimation. In order to generate an appropriate dataset for the training of the neural networks, an experimental setup is presented, which is used to record the vibration data of the gearboxes. It becomes apparent that especially vibration data in the frequency domain is suitable for the estimation of the rotational speed. Furthermore, it could be shown that low sampling rates result in an inaccurate speed estimation. Likewise, signal lengths that are too short result in inaccurate estimates. Overall, the speed can be estimated with an accuracy of 7.88 rpm at an average speed of the test data set of 202 rpm.

The present work investigates the size of gear damage required for significant recognisable change in the vibration signal and presents a method to determine digital filter limits in order to emphasise the vibration behaviour in the time domain. For this purpose, two gears are artificially damaged to four different degrees. The damage levels are determined by a tactile gear measurement and the gears are inserted into two intact gearboxes. Measurements at different speeds are used to generate a representative dataset. On the one hand, the recorded signals are examined via cross-correlation in the time domain. On the other hand, the occurring frequency components are examined using a windowed fast Fourier transformation. Based on the two observations, a statement is made about the recognisability of the damage levels of the two gears in the vibration signal. Furthermore, smoothed spectra are calculated via linear prediction coefficients (LPC) and an appropriate number of required coefficients is estimated via the Akaike information criterion. Subsequently, the calculated prediction coefficients are used as coefficients of an all-pole filter to calculate difference spectra. Based on the difference spectra, filter limits for a digital filter are derived to emphasise the damaged tooth meshing in the time domain.