Amongst the six proposals, the reviewer selected the following to receive funding. Two of them are revised versions of project ideas handed in again.
- Learning-Aided Real-World Semantic Exploration
By Angela Schöllig, CIT and Stefan Leutenegger, CIT
The goal of this 6-month project will be to combine our expertise in visual-inertial SLAM (VI-SLAM), dense and semantic mapping, as well as safe navigation and control, to achieve real-world exploration with a drone at an unprecedented scale, robustness, and speed. To achieve this, we will be tightly integrating probabilistic VI-SLAM with mapping and control, and furthermore leverage learning-based completion and semantic understanding for efficient and safe exploration. The drone platform as well as some core algorithms have been developed by the teams, which makes this project feasible in the short timeframe of 6 months.
- EMIRAS, Ergonomic Master console for Intuitive Robot-Assisted Surgery
By Mahammad Ali Nasseri, School of Medicine and Klaus Bengler, School of Engineering and Design
Within this project, the laboratory of Medical Autonomy and Precision Surgery (MAPS) at the eye clinic and the chair of ergonomics (LfE) will join their expertise to address one of the most important challenges in microsurgical robotic frameworks, which is having an intuitive interface between the robot and the surgeon. Despite all the benefits, robotic-assisted surgery has not yet been broadly accepted by clinicians in most clinical fields due to a lack of intuitiveness and issues transferring the dexterity of the user to the robot. Consequently, for seamless integration of surgical robots into the clinical workflow, one of the major challenges is the user-centered design of intuitive, multi-domain (physical, visual, and auditory) interfaces for robot-assisted microsurgery, that allow for high precision control and dexterity. In this project, taking robot-assisted ophthalmology, as an immediate use case, we will design, prototype, and validate a novel master console concept for intuitive control in micro-robotic surgeries. The development will be done in an iterative design and testing process in close cooperation with surgeons and other operating room staff.
- BE-MINd, Bionic Limb Embodiment via Minimally Invasive Nanoelectrodes
By Christina Piazza, CIT and Kristen Kozielski, CIT):
We will investigate the use of nanoelectrodes to stimulate the SI cortex area corresponding to the hand. A first analysis will be conducted using noninvasive transcranial magnetic stimulation (TMS) to provide multisensory perceptual illusions. This will give important insights to design our injectable, wireless nanoelectrodes. These are made of 2 magnetoelectric (ME) materials, and thus allow us to wirelessly generate electric signals using an input magnetic field (Fig. 4). We hypothesize that this yields motor activation, plasticity between the SI and motor cortices, and ownership of an artificial limb. The work proposed within the scope of the Seed Funding will create a foundation for multi-step project with a larger perspective. Successful completion of the aims proposed below will found our collaboration, enable joint publications, and ideally, position us to seek additional funding.
- BuildingSwarmBot, The benefit analysis of swarm robotics on the construction site
By Johannes Fottner, School of Engineering and Design and Alois Knoll, Faculty of Informatics
Construction work is dangerous, physically demanding work in an unpredictable physical environment. In order to protect workers and meet European and global development targets, construction operations need to be drastically improved, while labor shortage and a low level of automation prevail as opposition. Despite the rise of research activities on so-called “single-task construction robots” (STCRs), the sightings of robots on construction sites and along with that the degree of automation is minimal compared to other industries. Potentials in safety, productivity and quality increase through increased automation are therefore enormous. Challenges to the adoption of STCRs in a swarm-like operating scenario prevail in all three sectors of machine intelligence: perception, artificial intelligence and robotics.
The funding will cover 50% of up to two (post)doctoral researchers for up to six months duration to initiate the new project. Successful projects can start immediately. MIRMI can support up to four innovative and ambitious ideas projects in each round. The next call will be in spring 2023. We will keep you posted.
