Efficient sensorimotor processing constrains an acting agent in its environment. Learning mechanisms allow the agent to extract the underlying correlations in sensorimotor streams exploiting possible sensory patterns and plausible motor actions. This research field focuses on the exploration of sensorimotor learning paradigms for embedding adaptive behaviours in robotic systems and demonstrates flexible control systems using neuromorphic hardware and neural-based adaptive control.

Autonomous Systems

Traffic simulation is a crucial tool for testing and evaluating hypotheses in the design of autonomous driving systems, where modeling and calibrating driver behaviour is key. In this research, we introduce a physics-informed approach to calibrating driver models by combining fuzzy logic and kinematics to produce a plausible model parametrization.

Industry Automation Technology

Predictive maintenance of industrial automation systems to reduce unplanned stops and maximize equipment life cycle. “If component A is showing over 20% baseline vibrations, while the temperature rises 0.5 degrees in component B, and the noise level rises with 10.6dB then it is likely that the doors will break in about 5 to 7 days.” The goal of the project is to learn normal operative parameters and sensory correlations in order to: 1) define limits for min/max ranges of critical parameters and 2) generate alarms for anomaly situations. We research neural network streaming algorithms capable of explaining the variance-covariance structure of a set of variables in a stream through linear combinations. The essentially neural algorithm is leveraged by novel incremental computation methods and systems operating on data streams and capable of achieving low-latency and high-throughput when learning from data streams, while maintaining resource usage guarantees for predictive maintenance tasks.

Traffic Control

Traffic congestion poses serious challenges to urban infrastructures through the unpredictable dynamic loading of their vehicular arteries. Despite the advances in traffic light control systems, there are still no straightforward solutions to the problem of optimal traffic signal timing. Fundamentally nonlinear, traffic flows exhibit both locally periodic dynamics and globally coupled correlations and uncertainty. In this project, we are developing an end-to-end system capable of modelling the cyclic dynamics of traffic flow and robustly compensate for uncertainty while still keeping the system feasible for real-world deployment. To achieve this goal, the system employs an efficient representation of the traffic flows and their dynamics in populations of spiking neural networks.


We conduct research and develop tools, models, and infrastructure needed to interpret large amounts of clinical data and enhance cancer treatments and our understanding of the disease. To this end, this research serves as a bridge between the data, the engineer, and the clinician in oncological practice. Thus, knowledge-based predictive mathematical modelling is used to fill gaps in sparse data; assist and train machine learning algorithms; provide measurable interpretations of complex and heterogeneous clinical data sets; and make patient-tailored predictions of cancer progression and response.


Virtual reality (VR) sensorimotor rehabilitation is still in its infancy, but will soon require avatars, digital alter-egos of patients' physical selves. Such embodied interfaces could stimulate patient perceptions in a rich and highly customised environment, where sensorimotor deficits, such as in chemotherapy-induced peripheral neuropathy, could be corrected. We work on neural network meta-learning systems exploiting the underlying correlations in body kinematics with potential to provide, within latency guarantees, personalised VR rehabilitation. The unsupervised meta-learners are able to extract underlying statistics from motion data by exploiting data regularities in order to describe the underlying manifold, or structure, of motion in cases of sensorimotor deficits.

Sports Technology

This research field in the SPICES lab proposes the development of a VR neural network controller for sport psychological (cognitive) and biomechanical training. By exploiting neuroscientific knowledge in sensorimotor processing, neural network-based learning algorithms, and VR avatar reconstruction, the consortium partners are developing an adaptive, affordable, and flexible novel solution for goalkeeper training in VR.