Other Contributions


In partnership with INCM – The Portuguese Mint and Official Printing Office, we created a new machine readable code, designed to be the next generation of codes, after QRCode. The main advantages are:
higher capacity, higher aesthetic value, can be read using a smartphone.


  • excise stamp for tobacco (already in production – start May/2019)
  • ID citizen card
  • Passports
  • Wine stamps

2 patents pending

Motivation: surveillance tasks can be dangerous in some scenarios; are dull and prone to be automated with teams of robots, hence multi-robot patrolling

Problem: multi-robot coordination to visit as often as possible security spots in large areas and clear threats

Sub-Problems: distributed decision (no central point of failure); resiliency and adaptation to varying team size


The use of sensorial stimulation has been applied in medical applications with great success. Its fusion with AR or VR aims at improving therapeutic applications but also enlarge the application field to industrial, teleoperation and telepresence applications.

AR, VR and interactive robots have to take into account the holistic nature of human senses and the situational, perceptive and cognitive context .

Brain-computer interfaces (BCI) (or brain-machine interfaces (BMI)) offer a direct communication channel between the brain and external devices, such as a computer or a robotic system. BCI is a promising research technology that has been mainly dedicated to restore communication, mobility and motor movement of people with severe motor disabilities. However, BCI research is rapidly widening its scope, including new medical and non-medical applications, such as neurorehabilitation (motor and neurodevelopment), augmented performance, education (learning), safety (driving assistance), and gaming.

Despite its enormous potential, BCI is still limited by low information throughput, reduced control dimensionality, low reliability, and high user’s mental workload, leading to low general usability of the BCIs. Therefore, BCI’s current properties still do not match the desired features that most applications require.

InSwitch is an ultra-compact, smart electronic device for performance improvement of line-operated, three-phase, induction motors. It is self-powered and can be installed inside the motor terminal box, measuring simultaneously the motor currents, voltages, and mechanical vibration. InSwitch can be used for motor soft-starting, dynamic load-based connection-mode change (leading to 5-15% energy savings in variable-load, fixed-speed motors), advanced motor protection (overload, overcurrent, stall, voltage/current unbalance, and phase loss) and fault diagnosis (combining motor current and vibration data), and remote control/monitoring (embedded wired and wireless communication). This novel device can potentially replace millions of traditional electromechanical contactors and allows the integration of induction motors into “Industry 4.0” data networks. InSwitch concept has been partially developed in ISR-UC, involving MSc and PhD students, and awarded with several national and international innovation prizes, including a 50 k€ funding from the Horizon 2020 – SME Instrument – Phase 1. It has two national patents.

It consists on a new method to automatically extract all fuzzy parameters of a Fuzzy Logic Controller (FLC) in order to control nonlinear industrial processes. The main objective is the extraction of a FLC from data extracted from a given process while it is being manually controlled.
In general, it is not easy to determine the most suitable fuzzy rules and membership functions to control the output of a plant, when the only available knowledge concerning the process is the empirical information transmitted by a human operator. Thus, a major challenge in current fuzzy control research is translating human empirical knowledge into FLCs. A possible candidate to meet this challenge is the application of the genetic algorithm (GA) approach to data extracted from a given process while it is being manually controlled.
Previous methods only optimize membership function parameters, considering the other components of the fuzzy system fixed, such as implication, aggregation and defuzzifier methods. Other common limitation is the selection of the correct set of input variables and delays. The variable selection process is usually either manual, or made by some auxiliary pre-selection criterion, and is not accompanied with the accurate selection of the right time delays, probably leading to low-accuracy results. A variable with the correct delay may contain more information about the output, than one which does not consider any delay. However, the variable and delay selection are not jointly performed with the learning of the fuzzy model, which precludes the global optimization of the prediction setting.

In the last decade there was strong progress in visual 3D reconstruction of urban landscapes with the 3D models being available in consumer applications such as Google Maps and Google Earth. These models enable visually pleasant fly-overs but, since they are obtained from satellite/aerial imagery, the quality of the experience strongly deteriorates whenever the user attempts walk-throughs at ground level. The ISR-UC devoted substantial effort to this problem by creating the StereoScan, which is a SfM pipeline that uses plane primitives instead of point primitives, and that outputs dense, visually pleasant urban models at ground-level from video acquired by stereo cameras mounted in a moving vehicle. More recently the team has built in its recent theory of using Affine Correspondences (ACs) for SfM to create the PiMatch that outputs piecewise planar reconstructions with the same quality as StereoScan but using monocular video. The PiMatch pipeline runs currently at 1 frame per second, which means that it can be potentially used for on-the-fly operation in robotic and autonomous driving applications. This line of work was awarded with a “Google Faculty Research Award” in 2014 and led to results that were published in high profile conferences (ECCV,CVPR) and journals (TPAMI).

Skid-steering platforms are robust and simple platforms commonly used in field robotics applications, including demining, agriculture, and environmental surveying and monitoring. These platforms are very energy-efficient and provide relatively accurate odometry estimates, while driving straight. However, their skid approach requires significant power to rotate and common odometry models are so inaccurate for this type of motion that heading estimates provided by this sensing source becomes useless.

In order to address these problems, the ISR Field Robotics team studied both the power consumption and the kinematics of skid-steered platforms and proposed new models that set themselves apart from existing work.

Having the ability to know the kinematic behavior of a robot during the design phase allows making proper decisions to achieve a desired kinematic behavior. During field operations, an accurate kinematic model allows navigation even under adverse conditions, like loss of GNSS satellites signals. The model proposed by our team derives a physical relationship connecting the geometry of a skid-steered to its kinematic model. Moreover, conventional approaches fail to take into account the changes in the kinematic behavior of a skid-steered robot as its center of mass changes. The change in the center of mass could be drastic, and hence a conventional kinematic model calibrated for one configuration would need to be re-calibrated for the new center of mass. The model proposed by our group explicitly takes into account the center of mass position and hence allows adjusting the kinematic model without any new calibration procedure.