In this paper, a novel extended permanent magnet synchronous motor model is presented that incorporates a quadratic flux-current function to represent the polarity-dependent saliency. The proposed model enables the design of sensorless polarity detection algorithms required by the initial position detection of permanent magnet synchronous motors. The novelty of the model is that it integrates the polarity-dependent saliency into the traditional machine model and introduces a new machine parameter, the polarity-dependent saliency coefficient, to specify the Hessian matrix of the flux-current function. A measurement method is presented for determination of the elements of the Hessian and the polarity-dependent saliency coefficient. The solution of the model is given for high-frequency sinusoidal pulsating voltage injection. Experimental results show that the proposed extended model accurately predicts the amplitudes and phases of the second harmonics of the motor currents, which are the carriers of the polarity-dependent information. This information enables a current measurement-based polarity detection algorithm using the phase difference between the fundamental and second harmonic of the apparent d-axis current. Both the presented measurement data and the proposed model show that injection in the d-direction is optimal for polarity detection.

In a vehicle, wheel speed sensors and inertial measurement units (IMUs) are present onboard, and their raw data can be used for localization estimation. Both wheel sensors and IMUs encounter challenges such as bias and measurement noise, which accumulate as errors over time. Even a slight inaccuracy or minor error can render the localization system unreliable and unusable in a matter of seconds. Traditional algorithms, such as the extended Kalman filter (EKF), have been applied for a long time in non-linear systems. These systems have white noise in both the system and in the estimation model. These approaches require deep knowledge of the non-linear noise characteristics of the sensors. On the other hand, as a subset of artificial intelligence (AI), neural network-based (NN) algorithms do not necessarily have these strict requirements. The current paper proposes an AI-based long short-term memory (LSTM) localization approach and evaluates its performance against the ground truth.

In this paper, we propose an extended permanent magnet synchronous motor model that incorporates a quadratic flux-current function to determine the polarity-dependent saliency and develop a sensorless initial position detection method that is capable of polarity detection. The quadratic model extension is based on the second-order Taylor polynomial of the flux-current function. We present the model equations in both the stator- and the rotor-oriented reference frames as well as the Park and inverse Park transforms of the quadratic terms and Hessian matrices. The transient behavior and the dependency of the extended model on the rotor position have been validated. The proposed initial position detection method injects non-modulated even square wave voltage signals and relies only on phase-current measurements. We analyzed the effect of the pulse length on the errors of the position estimation and polarity detection. We provide machine parameter-based approximations of the current components dependent on the rotor position and, based on them, present a quantitative pulse-length design method.

In the automotive industry, the reduction of development costs is of key importance. The development of electrical hardware is an expensive, time-consuming process with a lot of development stages (e.g., prototyping, electrical testing, mechanical testing, lifecycle testing). There is a growing need to increase the cost-effectiveness of the development and testing phases of embedded software using virtualization. Using this method, less prototype manufacturing is necessary since the simulations allow for faster and more effective discovery of a large portion of possible faults without building a hardware prototype. Renode is an open source embedded system simulation framework that facilitates software-based testing. The main goal of this paper is to explore the usability of the framework for automotive applications.

In this paper, we introduce a vehicle localization method designed for the SZEnergy race car, which competes in the Shell Eco-marathon. The proposed method comprises four different extended Kalman filter-based localization algorithms and a selection algorithm that determines the most suitable one based on vehicle speed, GNSS availability, and signal quality. The low-speed Kalman filters are based on a kinematic vehicle model while the high-speed variants are based on a dynamic vehicle model. Several measurements were performed during test maneuvers to evaluate the performance of the filters. The proposed method succesfully handles sensor miscalibration and GNSS outages.

In this paper, we propose a novel signal processing method for road surface roughness classification exclusively from wheel speed sensor signals. Road surface quality has a significant impact on fuel consumption and driving safety. Traditionally, it has been measured using specially equipped vehicles and, more recently, shared via cloud-based infrastructure; however, such data can be unavailable or quickly become outdated, making onboard solutions essential. We analyzed a large wheel speed sensor dataset from various test maneuvers to determine how road surface roughness influences spectral characteristics under different conditions, including changes in speed, tire pressure, payload, and tire type. The proposed road surface roughness classifier uses only wheel speed sensor signals. It selects signal segments appropriate for processing based on driving conditions and computes their order spectra. The number and relative power of the spectral peaks within the identified interval of interest of the order spectrum are related to road surface roughness. The implemented classifier is capable of distinguishing between rough and smooth road surfaces based on the number of peaks in the interval of interest. The overall accuracy of the implemented road surface roughness classifier was 87.4%.