Robotics: Multi-sensor Robots for Elderly Care (multiple projects)
The master projects outlined below are a part of the research project Multimodal Elderly Care Systems (MECS, financed by the Research Council of Norway, 2015-2019, IKTPLUSS program) with one PhD student and two postdoctoral researchers.
Overall project outline
The number of elderly people living at home is increasing and this trend is expected to continue. The challenge then would be how to provide technology that can handle the complex and different environments found in homes. Further, technology can easily be seen as a threat to privacy and lack of interpersonal contact. This project addresses these issues by user centered design of robotic systems and the development of adaptive technology. A part of this will be to demonstrate the benefits regarding both performance and privacy being improved by applying sensors like cameras on a robot companion rather than having them permanently mounted in a home. These would be used for detecting falls and other non-normal situations.
A general public introduction to the project (in Norwegian).
Using new sensor technology, we would also like to explore if it is possible to remotely monitor medical states like pulse, breathing etc. Rather than having elderly themselves activating their personal security alarm in the case of an emergency situation, a target of this project is to demonstrate automatic activation. Many systems for elderly have been designed but few have been adopted on a large scale. We think a key reason for this is limited user involvement and few iterations of user testing. Therefore, we will focus specifically on developing our systems with a large degree of user participation.
The planned platform to be used in the project is a robot companion (see picture) developed by Giraff in Sweden. It is an open platform and has for the last years been applied in a number of EU funded projects on elderly care. The main user interface is a screen, camera, speaker and microphone. Another alternative includes Turtlebot 2.
A number of robotics master projects are available related to the MECS project:
- Sensor-based navigation without a map. Navigation is one of the primary tasks for any robotic system which has received a considerable amount of attention, especially since the robots started becoming an important part of humans daily life. Despite remarkable abilities of modern robots in sensing and actuation, the problem of navigating in a complex environment is still a big challenge in almost all robotics applications. Moreover, this problem is relevant to other disciplines such as verification, computational biology, and computer animation. An algorithm to address this problem is said to be complete if it terminates in finite time, returning a valid solution if one exists, and failure otherwise. The motion planning problem is known to be very hard from the computational point of view. For example, a basic version of the motion planning problem, called the piano movers problem, is PSPACE-hard. In fact, while complete planning algorithms exist, their complexity makes them unsuitable for practical applications.To even make the problem more complicated, assume a mobile robot moving in an office without knowing the map of the environment. The only source of information is the readings of the robot’s sensory system such as cameras or range finders. Since the reading range of these devices is limited, designing a global navigation plan is impossible and the robot requires to move among surrounding obstacles towards the goal following an incremental local behavior. The purpose of this research project will be on designing efficient intelligent navigation methods for a mobile robot moving in a 2D bounded environment using different machine learning methods. A nice tutorial on motion planning is given in two papers by Steven. M. LaValle (part I and part II). For more details about different sensor-based navigation methods see the following papers. (paper 1, paper 2 and paper3)
- Multi-objective sampling-based motion planing. Sampling based motion planning is an interesting class of navigation algorithms where the planning occurs by sampling the configuration space (C-space). The main objective is to capture the connectivity of the C-space by sampling it. This randomized approach has its advantages in terms of providing fast solutions for difficult problems without requiring a detailed map of the environment. The drawback is that the solutions are widely regarded as suboptimal. Sampling based planners are not guaranteed to find a solution if one exists, a property that is referred to as completeness . They ensure a weaker notion of completeness that is probabilistic completeness . A solution will be provided, if one exists, given sufficient runtime of the algorithm (in some cases infinite runtime). However, optimality is the only fully studied goal of sampling-based methods and other potential planning objectives such as improving the navigation safety and reducing the processing time still remain as unsolved challenges. The main objective of this project will be to achieve different objectives in sampling-based motion planning without risking the performance inefficiency. This work includes simulation studies in MatLab or other suitable environments as well as experimental studies on real robotic systems such Turtlebot and UR5 robotic arm. For more details about sampling-based motion planning, you can take a look at the this book authored by Steven M. LaValle. Furthermore, you can read this paper which is a detailed review on different aspects of sampling-based motion planning, or these three papers (paper 1, paper 2 and paper 3) for more information about these algorithms.
Benefits of taking master project related to an externally funded research project:
- close collaboration with doctoral students (PhD) and postdoctoral researchers working on related topics
- strong focus on progressing international state-of-the-art research and publication in leading international journals and conferences (beneficial for later application to PhD or researcher positions)
- funding available for publishing results available at international conferences
Competences relevant for jobs in industry: programming of embedded systems, system modeling, prototyping of sensor/mechanical systems, smartphone app development, experience from an upcoming application area (service robots, interactive musical systems, elderly care robots,…)
- Robotics and Intelligent Systems group (coordinator)
- DESIGN group (IFI)
Oslo Municipality (Oslo kommune, Gamle Oslo)
Norwegian Centre for Integrated Care and Telemedicine (Tromsø)
XCENTER AS (3D sensor)
Novelda AS (ultra wideband sensor)
University of Hertfordshire
University of Reading Whiteknights