Peter Slaets - Academia.edu (original) (raw)

Papers by Peter Slaets

This paper describes a method to predict the energy consumption of dual driven electric vehicles ... more This paper describes a method to predict the energy consumption of dual driven electric vehicles (EVs) over a predefined trajectory. This way the voltages that need to be applied on the motors are calculated to optimize the trip for minimal energy consumption and traveling time. Since traveling time and energy consumption are conflicting constraints, the goal is to decrease energy consumption, without entailing a drastic increase in traveling time. An algorithm was designed to get a weighted optimization for these two constraints, taking into account the trajectory and the aerodynamics of the vehicle. For the aerodynamics also wind velocity and direction are taken into account, combined with the heading of the car. Concerning the trajectory, the route to follow is inserted as a constraint, and the slopes along this route and its environment are taken into account. Using this algorithm, control signals, which are the voltages applied on left and right motor of the vehicle, can be obtained, which make it possible to control the vehicle throughout the trajectory. The simulation shows that an increase in traveling time has a relatively bigger impact on energy consumption. If direction and wind velocity can be predicted, the algorithm can anticipate on a change in wind velocity, or wind direction. This way the vehicle will consume less energy. Lastly when the trajectory is free to choose by the algorithm, steep slopes will be avoided dependent on the defined weight factor.

Nature

I n January last year, a Japanese car ferry, the Soleil, became the first large vessel to navigat... more I n January last year, a Japanese car ferry, the Soleil, became the first large vessel to navigate without human intervention. The 220-metre-long ship automatically berthed and unberthed, turned, reversed and steered itself for 240 kilometres across the Iyonada Sea from Shinmoji in northern Kyushu-manoeuvres that even skilled human operators find challenging. It is early days, but ships are increasingly deploying sensors and artificial-intelligence (AI) systems to navigate, steer and avoid collisions. As with cars, such advances should improve safety, increase efficiency andalong with cleaner fuels and engines-reduce environmental impacts. This is crucial: 80% of global trade (around 11 billion tonnes) is transported by sea each year 1. In 2018, shipping generated around 3% (about 1,000 million tonnes) of global carbon dioxide emissions 2. The International Maritime Organization (IMO) has committed to halving the industry's greenhouse-gas emissions by 2050. Seafaring is risky and workers are in short supply. Inefficiencies and congestion at ports add delays and costs, as do accidents, such as the grounding of the container ship Ever Given in the Suez canal for six days in March 2021. Streamlining passage through locks, reducing energy consumption and negotiating manoeuvres to avoid collisions would enable safer and more optimal use of waterways. Some small, fully autonomous boats,

IEEE Transactions on Emerging Topics in Computing

The ability to identify and temporally segment finegrained actions in motion capture sequences is... more The ability to identify and temporally segment finegrained actions in motion capture sequences is crucial for applications in human movement analysis. Motion capture is typically performed with optical or inertial measurement systems, which encode human movement as a time series of human joint locations and orientations or their higher-order representations. State-of-the-art action segmentation approaches use multiple stages of temporal convolutions. The main idea is to generate an initial prediction with several layers of temporal convolutions and refine these predictions over multiple stages, also with temporal convolutions. Although these approaches capture longterm temporal patterns, the initial predictions do not adequately consider the spatial hierarchy among the human joints. To address this limitation, we present multi-stage spatial-temporal graph convolutional neural networks (MS-GCN). Our framework decouples the architecture of the initial prediction generation stage from the refinement stages. Specifically, we replace the initial stage of temporal convolutions with spatial-temporal graph convolutions, which better exploit the spatial configuration of the joints and their temporal dynamics. Our framework was compared to four strong baselines on five tasks. Experimental results demonstrate that our framework achieves state-of-the-art performance.

Packaging Technology and Science

BMC Medical Informatics and Decision Making

BackgroundAlthough deep neural networks (DNNs) are showing state of the art performance in clinic... more BackgroundAlthough deep neural networks (DNNs) are showing state of the art performance in clinical gait analysis, they are considered to be black-box algorithms. In other words, there is a lack of direct understanding of a DNN’s ability to identify relevant features, hindering clinical acceptance. Interpretability methods have been developed to ameliorate this concern by providing a way to explain DNN predictions.MethodsThis paper proposes the use of an interpretability method to explain DNN decisions for classifying the movement that precedes freezing of gait (FOG), one of the most debilitating symptoms of Parkinson’s disease (PD). The proposed two-stage pipeline consists of (1) a convolutional neural network (CNN) to model the reduction of movement present before a FOG episode, and (2) layer-wise relevance propagation (LRP) to visualize the underlying features that the CNN perceives as important to model the pathology. The CNN was trained with the sagittal plane kinematics from a...

Journal of NeuroEngineering and Rehabilitation

Background Freezing of gait (FOG) is a common and debilitating gait impairment in Parkinson’s dis... more Background Freezing of gait (FOG) is a common and debilitating gait impairment in Parkinson’s disease. Further insight into this phenomenon is hampered by the difficulty to objectively assess FOG. To meet this clinical need, this paper proposes an automated motion-capture-based FOG assessment method driven by a novel deep neural network. Methods Automated FOG assessment can be formulated as an action segmentation problem, where temporal models are tasked to recognize and temporally localize the FOG segments in untrimmed motion capture trials. This paper takes a closer look at the performance of state-of-the-art action segmentation models when tasked to automatically assess FOG. Furthermore, a novel deep neural network architecture is proposed that aims to better capture the spatial and temporal dependencies than the state-of-the-art baselines. The proposed network, termed multi-stage spatial-temporal graph convolutional network (MS-GCN), combines the spatial-temporal graph convoluti...

Journal of Transportation Technologies

This paper assesses the vehicle dynamics of a new cargo bike concept developed for euro pallet si... more This paper assesses the vehicle dynamics of a new cargo bike concept developed for euro pallet sized cargo. The cargo bike developed is for last-mile delivery. Different aspects of manoeuvrability and stability are examined using a series of manoeuvres based on tests from the automotive industry combined with bicycle industry regulations. These manoeuvres objectively evaluate and determine the handling capabilities of the cargo bike concept. Those tests can be compared using the results of the benchmark vehicles. The results conclude the new cargo bike has proper vehicle dynamics above the majority of benchmark vehicles but there is still room for improvement.

Transport Problems

In many city centres and some large urban areas, access restrictions are imposed on trucks and va... more In many city centres and some large urban areas, access restrictions are imposed on trucks and vans. On the other hand, the number of e-commerce packages to be delivered daily is increasing rapidly, and service companies of all kinds also have to serve their customers in these areas. The use of cargo bikes is seen as a possible solution. The direct reason for this research is the observation that there is still room for a new type of cargo bike that meets the needs of the outlined target groups. This article summarises how the quality function deployment (QFD) method has been used for the systematic development of an electric cargo tricycle that meets these needs. The developed tricycle distinguishes itself from existing cargo bikes mainly by its loading capacity, stability and manoeuvrability. The prototype tests performed by professional couriers were so positive that a pilot series was built.

Additional file 1. Table S1: The evaluated hyperparameter space of the convolutional neural netwo... more Additional file 1. Table S1: The evaluated hyperparameter space of the convolutional neural network (CNN).

2021 6th International Conference on Mechanical Engineering and Robotics Research (ICMERR), 2021

In recent years, Simultaneous Localisation and Map-ping (SLAM) in dynamic environments received m... more In recent years, Simultaneous Localisation and Map-ping (SLAM) in dynamic environments received more and more attention. Most approaches focus on efficiently removing dynamic objects present within the scene to perform SLAM with the assumption of a static environment. Some approaches incorporate dynamic objects within the optimization problem to perform SLAM and dynamic object tracking concurrently. In this paper, we propose to incorporate information from dynamic objects into a 2D graph-based SLAM approach. We experimentally show that, by adding a measurement function of the dynamic objects to the front-end graph structure, and adopting a motion model of the object, the trajectory of the dynamic object as well as the robot's trajectory can be substantially improved in the absence of static features within the graph. Experimental results based on simulated data and data from a differential drive robot with a LiDAR sensor validate this approach.

OCEANS 2019 - Marseille, 2019

This study discusses the experimental identification of the dynamic characteristics for an operat... more This study discusses the experimental identification of the dynamic characteristics for an operational 1/25 scale model of the Watertruck+ self-propelling barge of CEMT class I. Parameters such as flow rate, acceleration, thrust and speed-dependent resistance forces in the longitudinal direction are determined. The operational scale model discussed in this paper includes a hull with length 1.54 m, width of 0.2 m and height of 0.2 m. All experiments are performed at an indoor pool to reduce environmental disturbances such as wind and current. The vessel’s behavior and propulsion system, i.e. a 4-pump system, is analyzed and subsequently validated by modeling the pump circuit and corresponding pipe losses. The experimental results of the resistance forces are validated by applying Computational Flow Dynamics in SolidWorks, using the Finite Volume Method. By generating the pump and pipeline characteristic, the operating point is determined that provides insight on the overall efficiency, which is equal to 11.36%. Moreover, a second order polynomial is fitted to the experimental resistance forces, with a correlation coefficient of 0.99, which suggest a high accuracy and feasibility of the identified parameters. Results indicate that minor hardware improvements in the pump circuit would improve the performance significantly; however, the current vessel and propulsion system are suited for envisioned future research objectives such as testing new control algorithms.

Sensors, 2022

Visible light positioning is one of the most popular technologies used for indoor positioning res... more Visible light positioning is one of the most popular technologies used for indoor positioning research. Like many other technologies, a calibration procedure is required before the system can be used. More specifically, the location and identity of each light source need to be determined. These parameters are often measured manually, which can be a labour-intensive and error-prone process. Previous work proposed the use of a mobile robot for data collection. However, this robot still needed to be steered by a human operator. In this work, we significantly improve the efficiency of calibration by proposing two novel methods that allow the robot to autonomously collect the required calibration data. In postprocessing, the necessary system parameters can be calculated from these data. The first novel method will be referred to as semi-autonomous calibration, and requires some prior knowledge of the LED locations and a map of the environment. The second, fully-autonomous calibration pro...

Indoor positioning is a challenging research topic. Over the years, many different measurement pr... more Indoor positioning is a challenging research topic. Over the years, many different measurement principles and algorithms have been proposed. Each system has its own advantages and drawbacks, therefore trade-offs have to be made. For example, one generally needs to make a trade-off between cost and accuracy. However, recent developments in sensing technology have led to commercial systems that advertise sub-decimeter positioning accuracy for less than e1k. In this paper, we benchmark the accuracy of indoor positioning systems by Pozyx labs and Marvelmind robotics, as well as the VIVE tracker by HTC and Aruco Marker tracking in OpenCV. Results show that these systems achieve an average dynamic positioning accuracy of approximately 150 mm, 20 mm, 8 mm and 100 mm, respectively. Keywords–Indoor positioning; benchmarking; accuracy

Gait analysis is one of the most useful tools for assessing age-related conditions. This study de... more Gait analysis is one of the most useful tools for assessing age-related conditions. This study describes the preliminary validation of a novel vision-based method for unobtrusive, ambulatory monitoring of spatiotemporal gait parameters. The method uses a mobile platform that is equipped with a Microsoft Kinect. A proprietary, generative tracker is used for measuring the 3D segmental movement of the subject. A novel method was developed for extracting gait parameters from the raw joint measurements by using the relative distance between the two ankle joints. The results are assessed in terms of mean absolute error and mean absolute percentage error with respect to a motion capture system. The mean absolute error ± precision was 5.5 ± 3.5 cm for stride length, 1.7 ± 1.3 cm for step width, 0.93 ± 0.44 steps/min for cadence, and 2.5 ± 2.0% for single limb support. While these results are promising, additional experiments are required to assess the repeatability of this approach.

This paper presents a hardware and software architecture for vision-based human-robot interaction... more This paper presents a hardware and software architecture for vision-based human-robot interaction designed for field programmable gate array (FPGA) based embedded system. The configurable logic and memory blocks connected through programmable interconnects on the FPGA permit programmers to create complex systems running multiple processing cores in parallel, which motivated the authors to implement multiple vision algorithms and robot controllers on a single system-on-chip board, aiming at low cost, low power consumption and high performance in human-robot interaction for industrial and educational robots. The architecture is a product derived from a component based programming model together with a systematic methodology for concurrent system design proposed by the authors.status: publishe

2018 OCEANS - MTS/IEEE Kobe Techno-Oceans (OTO), 2018

Inland vessels play a vital role in inland freight transportation, nevertheless, they currently s... more Inland vessels play a vital role in inland freight transportation, nevertheless, they currently suffer from the competition by road and rail transport. Therefore, augmenting their motion control - with the future possibility of unmanned barges in mind - could increase their competitiveness and induce a paradigm shift in the inland freight transport sector. However, there has been relatively little analysis of the manoeuvring characteristics of small inland vessels or self-propelled barges. Consequently, this study uses Computational Fluid Dynamics (CFD) to identify a single-variable second-order surge resistance model for a typical European Class (CEMT) type I vessel. This identification method applied the OpenFOAM software to solve the Reynolds Averaged Navier Stokes (RANS) equations by deploying a Volume of Fluid (VOF) approach and a k-omega turbulence model. This CFD methodology was first benchmarked on a KVLCC2 hull of which experimental data are publicly available to validate t...

Journal of Lightwave Technology, 2020

Visible light positioning has the potential to be a cost-effective technology for accurate indoor... more Visible light positioning has the potential to be a cost-effective technology for accurate indoor positioning. However, existing approaches often require large amounts of incoming data, usually in the form of high resolution images or dense lighting distributions. Additionally, a line of sight between transmitter and receiver is generally required at all times. In this work, we present a positioning approach that combines measurements from a camera, encoders and a gyroscope. We compare multiple algorithms for fusing these data, namely an extended Kalman filter, a particle filter and a hybrid approach. The end result is a system that provides location estimates even with sparse lighting distributions and temporary outages, yet achieves an average accuracy of 2 to 4 cm. Even in the 95th percentile of the cumulative error distribution, accuracy can be as low as 2 cm and is often lower than 10 cm. Moreover, due to the use of a low-resolution camera (640x480 pixels) and efficient fusion algorithms, the latency is relatively low on a standard laptop (between 5.6 and 21 milliseconds). Even on a low-cost embedded board, latency generally does not exceed 100 milliseconds. We validate our approach experimentally and show that it is robust under a wide range of illumination conditions.

Robotics in Education, 2019

Teaching robotics to engineering students can be a challenging endeavor. In order to provide hand... more Teaching robotics to engineering students can be a challenging endeavor. In order to provide hands-on experiences, physical robot platforms are required. Previously, obtaining these platforms could be expensive, and required a lot of technical expertise from teaching staff. However, more recent models address these issues, therefore providing more opportunities for hands-on sessions. In this paper, we describe how we used the Turtlebot 3 mobile robot in master courses at KU Leuven. We provide an overview of the main functionalities, and suggest a number of improvements to further lower the learning curve for students. Additionally, we elaborate on the curriculum and learning outcomes of two courses that utilized Turtlebots in practically oriented sessions.