Walter Daems professor

Research Walter Daems

Research team

Expertise

Digital Signal Processing, Embedded systems for sensing and actuation, Prototype design in industrial and healthcare applications.

Situational aware navigation and mapping (SITANAV). 01/04/2024 - 31/03/2028

Abstract

SITANAV aims at increasing the reliability and robustness of Autonomous Vehicles (AV) operating in logistic, industrial and agricultural facilities by focusing on the following problems raised by companies in the consortium: 1) the limited flexibility of AVs in complex environments, 2) the high costs when localization requires additional infrastructure, 3) the high deployment and layout reconfiguration costs, and 4) the high memory footprint of discrete metric maps that hinders applications in large environments. SITANAV's key idea to overcome these problems is to increase the situational-awareness in navigation and mapping capabilities of autonomous vehicles. This will provide AVs with higher levels of self-adaptation based on the current situation through explainable decision-making via semantic maps and reasoning. There are three main technological barriers that SITANAV has to overcome: 1) lack of models fordescribing situations, 2) lack of capabilities to reason about objects and maps, and 3) lack of situational awareness in AVs' decision-making for navigation. The approach to remove these barriers is a framework that combines a metric-semantic map with situational models, which describe a set of relations that connect an AV's motion and perception capabilities to a particular situation. For example, when the AV can find the appropriate pieces of information to infer the current situation from the perceived environment, it can select and configure its perception and control behaviors (situational aware decision-making) to achieve the desired robustness and performance for the application at hand (e.g., detecting the situation of a partly blocked pathway and switching to a narrow-space navigation). The proposed method will extend existing graph-based models and tools with new features, reasoning, and query answering mechanisms, to gradually increase AVs' situational assessment capabilities. The improvements will be in small-scale iterations, following a continuous integration approach. This will be accomplished by two running demos (indoor and outdoor) with increasing complexity throughout the project. The SITANAV models and software are designed with forward compatibility in mind, because we can now already foresee many future extensions, such as new types of semantic features, memory and learning capabilities, and the integration of task planning.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

Conditioned harsh outdoor environment for perception systems of autonomous applications (CAVE). 01/02/2024 - 31/01/2026

Abstract

Autonomous mobile systems might fail for many reasons, but one of them is when the harshness of the environment increases. It is difficult for OEMs, integrators, sensor and hardware components providers to design a robust autonomous mobile system based on traditional testing methods. Especially perception systems are challenged in realistic and relevant harsh conditions (e.g. rain, fog, direct sunlight). Currently, testing of perception systems is done by waiting for these conditions to happen in real-life – which can easily cost weeks of waiting. When an update is done on the hardware of the perception system (e.g. a coating on the lens is added) the exact same test is needed to verify an improvement. However, in real-life this exact same harsh condition cannot be reproduced. So, there is a need for a modular, validated testing facility that allows controllable and measurable conditions, to enable repeatable and controlled harsh conditions. CAVE_INFRA aims to develop a fixed perception test facility which can control and measure rain, snow, fog, illumination, dust and debris conditions, including its digital twin and a real-life validation. We aim to provide the following services: i) Sensing hardware (incl coatings/cleaning systems) and software performance evaluation in harsh conditions, including benchmarking to support sensor selection ii) Harsh condition model and/or sensor model derivation iii) Training or validation of AI models for objects / human detection and pose estimation iv) Degradation tests in harsh conditions v) Generate test data and scenarios that can be used for driving out own research but also for certification purposes and discussions with certification bodies such as TuV. To produce the harsh conditions in realistic scenarios, there are different actuation systems foreseen to respectively actuate the perception system under test, the target objects to be detected, and some of the generated conditions such as diverse illumination systems to create dynamic contrast.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

ABN HaFreeS MVP Prototype. 01/01/2024 - 31/12/2024

Abstract

In this project we investigated during the first phase (IOF-POC ABN HaFrees Feasibility) whether it was possible to develop a hands-free kit for bicycles. The main goals here were ease of use and call quality at speeds above 10 to 15 km/h. To this end, we mapped the market, from which it became clear that the first focus should be on the functional user (i.e. the professional who wants to use his work-related commute by bike to call colleagues, customers, etc.). The first tests showed that a significant suppression of wind noise is possible using several techniques (selection of good microphones, the choice of an optimal arrangement of each individual microphone combined in an optimal configuration, appropriate shielding i.c.w. the right signal processing algorithms). The techniques on their own do not provide sufficient improvement, but the delta is sufficient so that the combination should allow for a quality conversation at 25km/h. In this second phase of the project, we want to develop a minium viable product (MVP) prototype, which should allow to (1) characterize the product on its main qualities, (2) benchmark the product against competing products, (3) set up tests in view of user feedback and (4) define the further direction of the valorization. An essential hurdle here is the intellectual protection of the technology.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

Hybrid AI for Predictive Road Maintenance (HAIRoad). 01/10/2023 - 30/09/2025

Abstract

The current approach to monitoring road quality is based on manual inspections and is labor intensive and relatively expensive. Hybrid AI for Predictive Road Maintenance (HAIRoad) aims to use (hybrid) AI to map the condition of the road network and make recommendations for road maintenance. An efficient and robust data pipeline will be developed using MLOps tools, which allow easy switching between model development and implementation/production. Three demonstrators will illustrate the feasibility of the approach: one with the Port of Antwerp Bruges and two at the municipal level. The demonstrators will allow to validate both the more technical aspects and the market potential. HAIRoad will deliver several innovations such as automated detection of the road conditions, new indicators for road management, sensor fusion by combining information from multiple sensors, and the application of hybrid-AI where we will incorporate physical models of road degradation into data-driven machine learning models.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

Natural objects rendering for economic AI models (NORM.AI). 01/09/2023 - 31/08/2027

Abstract

Natural objects (vegetables, fruits, food, etc.) are omnipresent in different industrial applications: food sorting, vegetable spray treatments, precision & automated farming, etc. Automating these applications to deal with large variabilities of natural objects (object's detection, recognition, pose estimation, etc.), requires innovative technologies that are enabled by Artificial Intelligence (AI) that has the ability to generalize to variabilities. However, training these AI models would require thousands of images / videos with detailed annotations of different items. In the state of the art, one needs >10k images to (re-)train an AI model with an accuracy of >90%, when, in average one minute is needed to annotate one 'real' image, however these can increase drastically depending on the use case at hand and the variability around it. The more variability one wants to cover, the more training images are needed. These findings clearly indicate that in order to be able to deploy AI models in the industrial applications, innovative techniques are highly needed to remove the burdens of data annotations2. These techniques need also to be easily usable by end users to avoid large amount of manual work to update the proposed methodology to new applications. NORM.AI builds further on the successful results from PILS SBO3,4, where rendering techniques were applied to industrial products with CAD (Computer Aided Design) information, to retrieve AI (synthetic) training data from updated CAD with radiance models. While CAD facilitates synthetic data generation in PILS SBO by providing a reference model to start rendering from, the goal of NORM.AI project is to extend this research to Natural objects where no CAD is available. Therefore, defining a reference model to start rendering from, is part of the research in the project. Creating variations from the reference model that takes both spatial & time changes of the natural objects and the natural scenes, as well as finding a sweet spot between real data augmentation techniques & synthetic data generation techniques constitute another research challenge in the project. This research will allow to identify economic scenarios of training data generation, taking into account their effect into AI model's accuracy and robustness. The project focuses into three research applications: 1- Food sorting applications, where 2D images are used to detect & sort fruits & vegetables, as they are coming, for example, in a conveyor system. 2- Crop monitoring applications, where images from 2D cameras, for example, installed in a harvester, are used to detect vine's rows, crop distribution, etc. 3- Weed monitoring applications, where 2D images guide a spraying system to locally sprayweeds in a high precision.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

Strengthening the capacity for excellence of Slovenian and Croatian innovation ecosystems to support the digital and green transitions of maritime regions (INNO2MARE). 01/01/2023 - 31/12/2026

Abstract

The main goal of INNO2MARE is to strengthen the capacity for excellence of Western Slovenian and Adriatic Croatian innovation ecosystems through a set of jointly designed and implemented actions that will support the digital and green transitions of the maritime and connected industries. Based on an in-depth mapping of the ecosystems and needs & gaps analysis, the consortium will formulate a long-term R&I strategy aligned with regional, national and EU strategies, as a visionary framework, and a joint action & investment plan, with concrete steps for building coordinated, resilient, attractive and sustainable maritime innovation ecosystems. To support the joint strategy and provide a model for the future collaborative R&I of the ecosystems' actors, the project will implement three R&I pilot projects that address some of the key challenges related to maritime education and training, security & safety in marine traffic as well as energy conversion and managementsystems' efficiency. These pilots will be the basisfor further development,scale-up and translation of the generated research results into innovative business opportunities through the coordinated mobilisation of public and private funding. The consortium will also implement innovative programmes that will support the engagement of citizens in the innovation processes, knowledge transfer for mutual learning, entrepreneurship & smart skills training and attraction of best talents, involving more than 1.000 participants across the Quadruple Helix. In all the project activities, the two ecosystems will strongly benefit from the sharing of best practices of the Flemish ecosystem, one of the most developed maritime innovation ecosystems globally. The project will contribute to reducing the innovation divide in Europe by systematically connecting the innovation actors within and between the ecosystems and creating synergies in R&I investments' planning and execution, thus developing a true innovation culture

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

Nexor - Cyber-Physical Systems for the Industry 4.0 era 01/01/2021 - 31/12/2026

Abstract

The fourth industrial revolution (Industry 4.0 as it is commonly referred to) is driven by extreme digitalization, enabled by tremendous computing capacity, smart collaborating machines and wireless computer networks. In the last six years, Nexor — a multi-disciplinary research consortium blending expertise from four Antwerp research labs — has built up a solid track record therein. We are currently strengthening the consortium in order to establish our position in the European eco-system. This project proposal specifies our 2021 - 2026 roadmap, with the explicit aim to empower industrial partners to tackle their industry 4.0 challenges. We follow a demand driven approach, convincing industrial partners to pick up our innovative research ideas, either by means of joint research projects (TRL 5—7) or via technology licenses.

Researcher(s)

Research team(s)

Project website

Project type(s)

  • Research Project

Dotation for the structural collaboration with Flanders Make. 01/01/2021 - 31/12/2024

Abstract

Flanders Make's mission is to strengthen the international competitiveness of the Flemish manufacturing industry on the long term through industry-driven, precompetitive, excellent research in the field of mechatronics, product development methods and advanced production technologies and by maximizing valorisation in these areas.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

Cyber-physical system for personalized intranasal drug delivery. 01/05/2023 - 30/04/2024

Abstract

The upper nasal space is a unique drug delivery environment and has long been underutilized. Nose-to-brain delivery of therapeutic agents for neurodegenerative disorders does not have the same limitations as other existing administration methods. However, intranasal drug delivery has its own set of challenges because of the nasal shape's heterogeneous character and dynamics. This large morphological variation demands a person-specific approach that enables live sampling of the nasal structure and subsequently uses it during administration of the therapeutic agent. This project is the first and important step in the creation an efficacious intranasal drug delivery solution: designing and building the portable device able to capture the three-dimensional nasal shape in a non-invasive manner.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

ABN HaFreeS Feasibility. 01/09/2022 - 31/12/2023

Abstract

In this project, we take the first steps in the development a novel hands-free communication set for use on bicycles. The main advantages of our solution in comparison to current solutions are call quality and convenience. Call quality is our main selling point: wind noise, traffic noise and contact noise impede comfortable calling at a speed above 10-15 km/h with the current available technology; we aim to overcome these shortcomings using technology building blocks available in Cosys-lab that have matured in other application domains. We mainly focus on showing technology feasibility, initiating a market study and perform initial user outreach activities. We will also start preparing the design of an MVP prototype. These activities are an essential first step to determine if it is worthwhile to pursue the end goal of commercializing the solution in a spin-off. If the technology is shown to work and the valorisation potential lives up to our expectations, we will undertake further steps in later projects to develop a deep market insight, a convincing MVP prototype and a solid value chain. These elements are necessary to reach the end goal of starting a VC-funding-free spin-off, bootstrapping with funding gathered in a crowdfunding campaign.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

Mixed absolute and relative localization (MARLOC). 01/04/2021 - 31/03/2023

Abstract

In this project, we will combine fixed infrastructure localization (markers, UWB) with relative localization perception (SLAM, odometry) to arrive at a highly accurate and robust location estimate for mobile robots. We will work on simulation models and performance prediction based on deep learning techniques to optimize the localization setup.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

ModAU - Modernized Auscultations for Remote Monitoring. 01/01/2021 - 31/12/2022

Abstract

In hospitals where there is a significant workload, performing auscultations can be a timeconsuming process, which also exposes the medical personnel to potentially contagious diseases. Current systems that allow remote auscultations are often not fit for use with large amounts of patients, long-term use, or are limited in terms of functionality. The major drawback in current remote auscultation systems is the relatively bulky acoustic coupler which makes part of the stethoscope assembly. This physical dimension reduces the applicability for long-term monitoring, because of the discomfort for the patient and the inherent risk of decubitus wounds. In this project, we will investigate the construction of thinner stethoscopes, increasing patient comfort

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

SmartFlush. 01/02/2020 - 31/07/2021

Abstract

In this project we develop smart flushing solutions together with our industrial partner, IPEE nv. We use advanced techniques to improve the processing of their proprietary sensor data. We also operationalize a deployement setup.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

A Hybrid SLAM approach for autonomous mobile systems (HySLAM_SBO) 01/07/2019 - 31/03/2024

Abstract

In HySLAM,we will investigate the introduction of semantics in SLAM. We will introduce new probabilistic models which are based on scene understanding to increase the conditioning of the SLAM problem. Taking into account the underlying dynamics of the objects, and their effect on the perceptual scene, can help to increase the robustness of the SLAM algorithms. We will demonstrate the efficacy of the algorithm in a 2D and 3D test case.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

AirLeakSLAM: On-line detection of pressured-air leaks in industrial environments using passive and active ultrasonic sensing. 01/05/2018 - 30/06/2019

Abstract

A large amount of energy is lost annually due to leaks in compressed air networks. The combination of SLAM and 3D-ultrasonic measurement techniques enables to automate the measurement and registration of these leaks without requiring manpower. Therefore, measurements can be conducted in a continuous (on line) instead of an incidentally manner. The goal of the project is to demonstrate the power and the opportunities of the system for the user of the compressed air system, and to further quantify the value creation opportunity.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

Localization system for accurate tracking and navigation for autonomous operation (LOCATOR_ICON) 01/11/2016 - 31/10/2018

Abstract

In order to choose the right combination and placement of sensors to perform sensor-fusion based indoor localization in industrial environments, a framework for designing systems for global and relative localization can facilitate the development. To quantify the performance of various sensors in this operational context, models of these sensors need to be developed. These models will be probabilistic in nature in order to be used with the aforementioned sensor fusion techniques and to calculate confidence intervals where safety is an issue. The sensor models will be parametrized and will be able to incorporate in-situ experimental measurements to make the simulations more accurate.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

Scale-free passive acoustic localization using a wireless synchronized sensor network. 01/10/2016 - 30/09/2020

Abstract

During this project we will develop a framework which allows passive localization of acoustic sources using a synchronized wireless sensor network. Synchronization of the wireless microphone array will be performed using a distributed synchronization scheme absent of a master time representation. The framework will support automatic calibration of the microphone array with minimal human intervention. The location estimate of the acoustic sources will be performed using a probabilistic localization algorithm in combination with known statistics about the behavior of the acoustic source. The framework will be virtually scale-free, which means that the sensor network can be used for tracking a wide variety of acoustic sources in a wide variety of application domains.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

Cost-effective vibroacoustic monitoring (vibmon_icon). 01/10/2015 - 31/12/2017

Abstract

The Cost effective vibroacoustic monitoring project will attempt to prove the technical and economic feasibility of cost effective vibroacoustic monitoring systems for continuous online condition and process monitoring of rotating machine elements in quasi stationary conditions. The project will make use of new opportunities enabled by the advent of cost effective sensors, like MEMS accelerometers, microphones, and microphone arrays, and cost effective embedded platforms that in combination can provide an efficient solution for continuous monitoring. The generic part of the project will assess the technical limitations of cost effective sensors compared with high-end ones and will overcome this limitations by develop novel digital signal processing algorithms for: • Automatic pre-processing and data cleaning of raw data recorded by cost-effective sensors in order to eliminate non-physical features present in the signals generated by certain cost effective sensors; • Feature extraction for fault detection and identification that can provide reliable diagnostic information and can deal the technical limitations of cost-effective sensors like limited bandwidth, high noise density, and lower sensitivity; • Online tachometer-less estimation of rotational speed in order to reduce the cost of the total solution by eliminating high precision speed sensors; • Reducing of the amount of data generated by the monitoring system while maximizing the amount of information to diminish the communication and data stream handling costs; The project will develop a technology validation platform for a cost effective vibroacoustic monitoring system including sensors, acquisition hardware, embedded processing unit and local digital signal processing software.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

Deriving sampling and memory strategies for 3D sonar systems to support electric wheelchair users 01/12/2014 - 30/11/2015

Abstract

This project represents a research agreement between the UA and on the onther hand IWT. UA provides IWT research results mentioned in the title of the project under the conditions as stipulated in this contract.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project

Service/Reuse for Assistive technology Delivery/design (SeRenADe). 01/01/2014 - 31/12/2017

Abstract

This project represents a formal research agreement between UA and on the other hand Gouverneur Kinsbergen Centre. UA provides Gouverneur Kinsbergen Centreresearch results mentioned in the title of the project under the conditions as stipulated in this contract.

Researcher(s)

Research team(s)

Project type(s)

  • Research Project