In this paper, one solution for an end-to-end deep neural network for autonomous driving is presented. The main objective of our work was to achieve autonomous driving with a light deep neural network suitable for deployment on embedded automotive platforms. There are several end-to-end deep neural networks used for autonomous driving, where the input to the machine learning algorithm are camera images and the output is the steering angle prediction, but those convolutional neural networks are significantly more complex than the network architecture we are proposing. The network architecture, computational complexity, and performance evaluation during autonomous driving using our network are compared with two other convolutional neural networks that we re-implemented with the aim to have an objective evaluation of the proposed network. The trained model of the proposed network is four times smaller than the PilotNet model and about 250 times smaller than AlexNet model. While complexity and size of the novel network are reduced in comparison to other models, which leads to lower latency and higher frame rate during inference, our network maintained the performance, achieving successful autonomous driving with similar efficiency compared to autonomous driving using two other models. Moreover, the proposed deep neural network downsized the needs for real-time inference hardware in terms of computational power, cost, and size.

Low Power Wide Area Networks (LPWANs) are gaining attention in both academia and industry by offering the possibility of connecting a large number of nodes over extended distances. LoRa is one of the technologies used as a physical layer in such networks. This paper investigates the LoRa links over seawater in two typical scenarios: clear Line-of-Sight (LOS) and obstructed path in two different Industrial, Scientific and Medical (ISM) radio bands: $868$ MHz and $434$ MHz. We used three different LoRa devices in the experiments: the Own Developed LoRa Transceiver (ODT) and two commercial transceivers. Firstly we investigated transceivers’ Receive Signal Strength Indicator (RSSI) and Signal-to-Noise (SNR) measurement chain linearity and provided correction factors for RSSI to correlate it with actual signal levels received at transceivers’ inputs. Next, we carried out field experiments for three different LoRa Spreading Factors, $SF\in \left[7,10,12\right]$ , within a bandwidth of $BW=125$ kHz and Coding Rate $CR=4/6$ . The experiments showed that LoRa links are fully feasible over seawater at distances at least $22$ km long, using only low-cost off-the-shelf rubber duck antennas in LOS path condition in both ISM bands. In addition, we showed that LoRa links can be established over $28$ km obstructed LOS oversea path in ISM $434$ MHz band, but using costly, higher gain antennas. Furthermore, the laboratory experiments revealed that RSSI is linear in a wide range, up to $-50$ dBm, whereas the SNR measurement chain goes into saturation for Received Signal Strength (RSS) values higher than $-100$ dBm. These findings enabled accurate interpretation of the results obtained in field experiments.

We present a novel design and implementation of a LoRa transceiver with enhanced characteristics which operates in two industrial, scientific, and medical (ISM) bands of 868 MHz and 434 MHz. Comparative tests with two widely utilized commercial transceivers were carried out to demonstrate the superior performances of the proposed transceiver, namely improved sensitivity by 15-25 dB and better immunity to interference. Field tests were performed by establishing a very long line-of-sight over land link of 112 km. The main achievements of this letter are: 1) the design of a novel high-sensitivity low-cost miniaturized LoRa transceiver; 2) demonstrating a very long LoRa link with low-cost commercial-off-the-shelf omni-directional rubber duck antennas with standard-compliant output power in the 868 MHz ISM band; and 3) temporal analysis of the link variability for different spreading factors (7, 10, 12) within a fixed LoRa bandwidth of 125 kHz.

Magneto-resistive and Hall angular position sensors are frequently used in servo systems, due to their reliability, longevity, and tolerance to environmental conditions. In this paper, a measurement error compensation is integrated into the servo system control structure. The compensation consists of two steps: a linear compensation of the magnetic field vector measurement error, and harmonic compensation of the angular position measurement error. The selfcalibration method uses data collected during rotation with maximal speed to identify parameters of the measurement process model. Identification is executed periodically, each time when the shaft makes a full rotation. The least squares method is utilized for identification of the linear compensation parameters, while the average rotational speed is used as an internal reference for the harmonic compensation parameters identification. The method, together with a fast convergence, increases the measurement accuracy by an order of magnitude. Experimentally achieved measurement error is below 0.7°, with residual variance below 0.05°, comparable with sensor differentiability.

The paper describes procedures and methods upon which a system has been developed for testing the running behavior of new or reconstructed electric locomotives prior to release them into regular service. Special attention has been paid to precise measurement of traction currents and voltages, as well as non-electrical quantities, such as torque, longitudinal force exerted at the wheel rim, angular and linear velocity, which vary in a wide range. In addition to choosing appropriate sensors and measuring transducers, specific interface circuits were designed and manufactured, and a laboratory model for measuring electrical and non-electrical quantities of traction electric motors was developed and realised. Based on the experience and the results obtained from the laboratory model, a computerised data acquisition system for testing the electric locomotives in the field was designed from such a computerised systems as application program for simultaneous multichannel measurement of mechanical and electrical quantities, and for off-line processing of measurement results. The systems were tested during the operation on measurement and analysis of tractive capabilities, quality of anti slip regulation system, and energy consumption of a modernized locomotive.