Robert Bosch (Taiwan)

Gadolinia-doped ceria (GDC) has been identified as a promising material for an application as a single-phase fuel electrode in solid oxide cells and receives interest for hydrogen operation [1-4] and CO2 electrolysis [5, 6]. The presence of both ionic and electronic conductivity as well as the electrocatalytic activity eliminates the need for an electronically conductive phase such as nickel and subsequently expands the active material-pore interface. But the rather high sintering temperature of GDC fuel electrodes, usually 1200 °C and above [2,4-8], results in a relatively coarse microstructure. Especially in low temperature applications, limited active surface areas cause significantly increased charge transfer reaction resistances.In this study, a physicochemically motivated impedance model based on a transmission line model [1, 3] is applied in order to optimize the electrode design of a single phase GDC fuel electrode that is contacted by a nickel current collector. The model is parametrized employing three different techniques as illustrated in figure 1 and is described in detail in Ref [3]. Details about the manufacturing of the GDC electrode can be found in [1].Four-point probe method is conducted in order to obtain the conductivity of GDC in a wide range of temperature und oxygen partial pressure (fig 1 a). Additionally, non-stoichiometry data [9] were used to separate the electronic and ionic conductivity and to calculate the chemical capacity of GDC. Focus ion beam tomography (fig 1 b) and electrochemical impedance spectroscopy (fig 1 c) were performed at four different GDC microstructures in order to reveal the microstructural parameters as well as the gas diffusion and charge transfer reaction impedance. The use of four distinct GDC microstructures for the parametrization and validation of the model ensures a high reliability for subsequent simulations.Figure 2 presents a simulation of the electrode impedance with different GDC microstructures. The microstructure is varied by simulating the microstructural parameters of GDC electrodes sintered at temperatures between 1300 °C and 1000 °C in 50 K steps. Two microstructures at the lowest and highest temperature are exemplary shown in (a) to visualize the coarsening towards higher temperatures. The simulation of the impedance and DRT spectra in (b) reveals an improvement of the polarization resistance by a factor of three from 1.5 Ω cm2 to 0.5 Ω cm2 at 600 °C and 97%/3% H2/H2O. The DRT spectra visualize the main impact in the lowest frequency peak allocated to the charge transfer reaction overlapped by minor gas diffusion [1, 3].In this contribution we demonstrate a physicochemically motivated impedance model that enables an optimization of the electrode design of a single-phase fuel electrode in a hydrogen steam operation at different temperatures. In addition to the entire microstructure, the model facilitates a modification of individual microstructural parameters including tortuosity, volume fraction, pore diameter or surface area. Geometrical parameters such as the electrode thickness or electrode area can be adjusted, too.

Testing real hardware and simulation models in combinationin a software- or hardware-in-the-loop set-up ischallenging. One of the key factors is the high demand foraccuracy in the simulation model. If classical modelingbased on physical principles is not sufficient to reach thedesired level of accuracy, hybrid modeling, the combinationof physical simulation models and machine learning can beapplied. In this publication, we train a hybrid model for acontrolled electric motor within the electro-hydraulicbraking system of a car under the conditions andrestrictions of a real engineering application in thefield. We apply state-of-the-art modeling patterns forthis, and further extend them with application specificmethodological optimizations. Finally, we investigate andshow the quantitative and qualitative advantages of theproposed approach for this specific application, resultingin a gain in accuracy by multiple factors.

Scene flow estimation is a foundational task for many robotic applications, including robust dynamic object detection, automatic labeling, and sensor synchronization. Two types of approaches to the problem have evolved: 1) Supervised and 2) optimization-based methods. Supervised methods are fast during inference and achieve high-quality results, however, they are limited by the need for large amounts of labeled training data and are susceptible to domain gaps. In contrast, unsupervised test-time optimization methods do not face the problem of domain gaps but usually suffer from substantial runtime, exhibit artifacts, or fail to converge to the right solution. In this work, we mitigate several limitations of existing optimization-based methods. To this end, we 1) introduce a simple voxel grid-based model that improves over the standard MLP-based formulation in multiple dimensions and 2) introduce a new multi-frame loss formulation. 3) We combine both contributions in our new method, termed Floxels. On the Argoverse 2 benchmark, Floxels is surpassed only by EulerFlow among unsupervised methods while achieving comparable performance at a fraction of the computational cost. Floxels achieves a massive speedup of more than ~60 – 140× over Euler-Flow, reducing the runtime from a day to 10× minutes per sequence. Over the faster but low-quality baseline, NSFP, Floxels achieves a speedup of ∼14×.

Bonsai, an ancient art form that blends horticultural precision with artistic expression, poses unique challenges to both beginners and seasoned practitioners due to its species-specific care requirements and intricate maintenance processes. This research introduces an innovative AI-based solution combining YOLOv8, an advanced object detection model, and Google Gemini LLM, to simplify bonsai species classification and care recommendations. The system is designed to accurately identify bonsai species and provide personalized, context-aware guidance on essential care aspects such as watering, pruning, and pest management, while dynamically adapting to environmental and seasonal variations. Featuring an intuitive interface with sub-100ms response times, the proposed solution enhances accessibility and engagement for users of all expertise levels. Testing affirms its capability to improve understanding, usability, and appreciation of bonsai cultivation. Beyond its application in bonsai care, this hybrid AI integration demonstrates the potential for broader adoption in fields requiring precise classification and tailored insights, showcasing a seamless blend of tradition and technology.

A high-power laser beam profiling system was set up to investigate the influence of the interaction between the laser beam and the process emissions during welding with a shaped beam profile. A positional instability of the beam on the workpiece in the order of magnitude of tens of µm and noticeable distortions of the beam shape were observed when no cross jet was used. Both perturbations were reduced when a cross jet was applied to remove the process emissions from the beam path and minimized when the cross jet was positioned closest to the workpiece.

Active safety systems for powered two wheelers (PTWs) are considered a key pillar to further reduce the number of accidents and thus of injured riders and fatalities. Enhanced awareness for the current riding situation is required to improve the performance of current systems as well as to enable new ones; this includes the detection of the rider’s intention – the action that is planned by the rider for the short-term future. The prediction of a continuous trajectory for the upcoming seconds of the ride is one way to express rider intention. Our work pursues the prediction of the PTW lateral dynamic state by means of a roll angle trajectory over the upcoming 4 s of riding. It thus considers the special vehicle dynamics characteristics of single-track vehicles that negotiate bends at a roll angle compared to cars. A deep learning (DL) prediction model that is based on a Long-Short Term Memory (LSTM) layer is optimized and trained for this task using a broad on-road riding dataset that focuses on the rural road environment. Inputs to the prediction model are PTW internal signals only, that are measurements of vehicle dynamics, rider inputs, and rider behavior. The latter two groups of signals are non-common for current series production PTWs and were especially added to our test bike before gathering the riding data set. The prediction performance of the optimized DL model is compared to a simple heuristic algorithm using multiple metrics in the roll angle and position trajectory domain. Evaluation on a representative test data set shows a significantly improved detection of rider intention by the DL model in all metrics. Reasonable lateral trajectory accuracy is achieved for at least 2 s of the total 4 s prediction horizon in 99 % of all test cases, given the chosen model architecture and input features. Furthermore, the feature importance of the especially added non-common measurement signals of steering and rider behavior is investigated in an ablation study. It reveals the importance of steering signals in the first second of the prediction horizon whereas the rider behavior signals aid trajectory prediction performance for up to 2.5 s.

Generalized Few-Shot Object Detection (G-FSOD) seeks to jointly detect base classes with abundant data and novel classes with limited data. Due to data scarcity, predictive uncertainties are more pronounced in G-FSOD than in conventional object detection. Unaccounting for these uncertainties leads to degraded overall detection performance and forgetting the base classes. However, previous G-FSOD works have not exploited these uncertainties. Upon examining the basic two-stage G-FSOD framework, which includes a Region Proposal Network (RPN) and a subsequent R-CNN, we observe that a straightforward integration of uncertainty estimation leads to detrimental performance. To this end, we first increase the model capacity by increasing the depth of the RPN and cascading multiple R-CNNs in an end-to-end manner. Next, we interleave the stages with uncertainty estimation and attention blocks. The aim is to progressively refine the proposals by exploiting the estimated uncertainties while attending to the discriminative features through the attention mechanism. Extensive experiments on the well-established G-FSOD benchmarks, MS-COCO and PASCAL-VOC, show that our proposed method sets a new G-FSOD standard.

<div class="section abstract"><div class="htmlview paragraph">Increasingly stringent tailpipe emissions regulations have prompted renewed interest in catalyst heating technology – where an integrated device supplies supplemental heat to accelerate catalyst ‘light-off’. Bosch and Boysen, following a collaborative multi-year effort, have developed a Rapid Catalyst Heating System (RCH) for gasoline-fueled applications. The RCH system provides upwards of 25 kW of thermal power, greatly enhancing catalyst performance and robustness. Additional benefits include reduction of precious metal loading (versus a ‘PGM-only’ approach) and avoidance of near-engine catalyst placement (limiting the need for enrichment strategies). The following paper provides a technical overview of the Bosch/Boysen (BOB) Rapid Catalyst Heating system – including a detailed review of the system’s architecture, key performance characteristics, and the associated impact on vehicle-level emissions.</div></div>

Purpose:Achieving greater health and social care integration is a policy priority in many countries, but challenges remain. We focused on governance and accountability for integrated care and explored arrangements that shape more integrated delivery models or systems in Italy, the Netherlands and Scotland. We also examined how the COVID-19 pandemic affected existing governance arrangements.Design/methodology/approach:A case study approach involving document review and semi-structured interviews with 35 stakeholders in 10 study sites between February 2021 and April 2022. We used the Transparency, Accountability, Participation, Integrity and Capability (TAPIC) framework to guide our analytical enquiry.Findings:Study sites ranged from bottom-up voluntary agreements in the Netherlands to top-down mandated integration in Scotland. Interviews identified seven themes that were seen to have helped or hindered integration efforts locally. Participants described a disconnect between what national or regional governments aspire to achieve and their own efforts to implement this vision. This resulted in blurred, and sometimes contradictory, lines of accountability between the centre and local sites. Flexibility and time to allow for national policies to be adapted to local contexts, and engaged local leaders, were seen to be key to delivering the integration agenda. Health care, and in particular acute hospital care, was reported to dominate social care in terms of policies, resource allocation and national monitoring systems, thereby undermining better collaboration locally. The pandemic highlighted and exacerbated existing strengths and weaknesses but was not seen as a major disruptor to the overall vision for the health and social care system.Research limitations:We included a relatively small number of interviews per study site, limiting our ability to explore complexities within sites.Originality:This study highlights that governance is relatively neglected as a focus of attention in this context but addressing governance challenges is key for successful collaboration.