Valeo (Germany)

Meeting EMC requirements on conducted emissions is one decisive development goal for modern e-drive systems. The well-known approach to achieve limits for 800 V systems via a filter comes with some drawbacks, like the increase of size or cost. Having those disadvantages in mind, the here proposed solution is to reduce the level of interference during operation by applying different modulation algorithms. Here the achieved improvements are quantified via a validated numerical computer model for an industrial automotive powertrain. As baseline for the comparison, a widely used modulation algorithm (Space Vector Modulation) is taken. Improvements between the investigated algorithms and the reference are shown for conducted emission results in frequency domain.

Recent cutting-edge methods for 3D object detection on point clouds are based on supervised learning methods. As these methods demand an extreme volume of data with the highest quality to train on, cost-effective annotation plays a crucial role in developing such perception algorithms, e.g., for autonomous vehicles or robots. Every inconsistency or error between the data captured by sensors and the subsequently generated labels might degrade the potential detection performance. Nevertheless, resources for annotation are usually very limited in terms of budget and time. We propose a straightforward yet highly effective technique called Click, Crop, and Detect (CCD) to address this issue. The core concept of CCD involves leveraging human input first to generate a prior rough localization of each object and employing 3D object detectors on a simplified cropped region of interest. We evaluate CCD across popular detectors such as PointPillars, CenterPoint, and TED on nuScenes and KITTI. Here, we show that only marginal changes to existing off-the-shelf detectors are required to make them compatible. Our method consistently outperforms state-of- the-art one-click detectors by 7.89% and 10.45% for cars and pedestrians, respectively, while being much more robust and precise on challenging, sparse inputs. This heavily increases label quality and efficiency when applied for semi-automated ground truth annotation.

This paper presents improvement potentials and noise mitigation techniques at the use case of ultrasonic distance ranging sensors for automotive applications. The key element of the product development for improved immunity behavior against noise sources in the low frequency band of below 200 kHz was the fusion of chip level noise quantification using system plausibility check features with the electromagnetic compatibility design optimization of electromechanical subcomponents. For this purpose, empirical and 3D simulation data are generated and analyzed. In addition, radiated immunity measurement methods are presented for the final component evaluation of the low frequency electric and magnetic immunity performance below 200 kHz.

Moving object detection and segmentation is an essential task in the Autonomous Driving pipeline. Detecting and isolating static and moving components of a vehicle's surroundings are particularly crucial in path planning and localization tasks. This paper proposes a novel real-time architecture for motion segmentation of Light Detection and Ranging (LiDAR) data. We use three successive scans of LiDAR data in 2D Bird's Eye View (BEV) representation to perform pixel-wise classification as static or moving. Furthermore, we propose a novel data augmentation technique to reduce the significant class imbalance between static and moving objects. We achieve this by artificially synthesizing moving objects by cutting and pasting static vehicles. We demonstrate a low latency of 8 ms on a commonly used automotive embedded platform, namely Nvidia Jetson Xavier. To the best of our knowledge, this is the first work directly performing motion segmentation in LiDAR BEV space. We provide quantitative results on the challenging SemanticKITTI dataset, and qualitative results are provided in https://youtu.be/2aJ-cL8b0LI.

3D object detection based on LiDAR point clouds is a crucial module in autonomous driving particularly for long range sensing. Most of the research is focused on achieving higher accuracy and these models are not optimized for deployment on embedded systems from the perspective of latency and power efficiency. For high speed driving scenarios, latency is a crucial parameter as it provides more time to react to dangerous situations. Typically a voxel or point-cloud based 3D convolution approach is utilized for this module. Firstly, they are inefficient on embedded platforms as they are not suitable for efficient parallelization. Secondly, they have a variable runtime due to level of sparsity of the scene which is against the determinism needed in a safety system. In this work, we aim to develop a very low latency algorithm with fixed runtime. We propose a novel semantic segmentation architecture as a single unified model for object center detection using key points, box predictions and orientation prediction using binned classification in a simpler Bird's Eye View (BEV) 2D representation. The proposed architecture can be trivially extended to include semantic segmentation classes like road without any additional computation. The proposed model has a latency of 4 ms on the embedded Nvidia Xavier platform. The model is 5X faster than other top accuracy models with a minimal accuracy degradation of 2% in Average Precision at IoU=0.5 on KITTI dataset.

Hairpin windings are widely used in automotive traction motors due to their high slot fill factor and their highly automated manufacturing process. The hairpins are arranged in layers in the stator slots, which offer additional degrees of freedom compared to conventional stranded wire windings in the design process. However, a circulating current within the parallel branches can occur, if the electromagnetic force or the effective branch resistance of each parallel branch differs. Consequently, several design rules have been formulated in the past. In this research, designs of short-pitched and full-pitched hairpin windings are presented and their excitation of circulating currents is evaluated. Time-transient finite element analyses of the current displacement effects in hairpin conductors in a single slot and in an exemplary machine are performed. It is shown that commonly applied design guidelines are not sufficient to fully prevent circulating currents. As a consequence, the definitions of weak and strong symmetry are suggested, the latter of which ensuring that circulating currents are not excited.

This paper provides a steady-state thermal characterisation of advanced insulating layers for isotropic electroconductive adhesive based Metal Core Printed Circuit Boards. Thermal measurements were conducted using the Thermal Transient Tester to analyse the properties of layers applied onto two different base materials - aluminium and copper. For each base material, four types of insulating layers were characterised. Size of the specimen plate for both base materials was selected as 100mm x 20mm × 1mm. The isotropic electroconductive adhesive was applied onto insulating layers to ensure electrical connection between seven SMD heat-generating components. These components are mounted in the identical distance to each other. The thermal characterisation of power SMD components and the assembling process based on isotropic electroconductive adhesive were evaluated. Application of advanced insulating layer was analysed to assess the thermal performance of complete MCPCB structure. It has been shown that the lowest thermal resistance of the aluminium based insulating layers is 21.1K/W, whereas the lowest thermal resistance of the copper based insulating layer is 15.5K/W.