• Research: University of Padova, Department of Information Engineering (UniPD)

  • Research: IT+Robotics S.r.l. (IT+R)

  • System Integrator: STAM S.r.l. (STAM)

  • End User: Institutul de Cercetari Electrotehnice (ICPE)



Figure 1: Robotic arm winding the coils directly on the stator

The team ITRXcell, composed by the University of Padova, IT+Robotics, STAM and ICPE, is proposing the FLEXICOIL project, aimed at developing a learning-based approach for robotized coils winding, to be used in the electric machines manufacturing industry. Wound coils can be found in several products (motors, generators, sensors) and for a wide range of applications. Smaller lot sizes and higher product flexibility cannot be achieved with conventional winding systems, because they are not flexible enough and manual labour causes high product costs. FLEXICOIL aims at overcoming this drawback by developing a robotic cell for coils winding. Three subsequent sets of activities will allow for developing a reconfigurable interactive manufacturing cell with learning capability, suitable to wind the coils of several kind of electric machines, basing on a simple teaching interface that can be easily used by operators without specific skills in robotics.

The solution proposed by FLEXICOIL will be affordable to a wide range of users: from small-medium enterprises (SMEs), producing small batches of motors and frequently changing product designs, to big companies, having a market request of several thousand standard units. This result will be possible by merging together ICPE’s experience in the electric motor field, the research skills of UniPD in the field of human-robot interaction and robot learning, the technology transfer and industrial vision capabilities of IT+R and the ability of STAM in integrating robotic and automation systems.

The project is divided in three main phases:

  • Phase 1 (freestyle): The free-style activities will aim to develop a solid learning system and a reliable human-robot interface. The robot will learn an arbitrary path through several pegs composing different possible routes and then pass a wire through this path, in order to demonstrate the consistency of the used approach. A number of subjects will teach the selected path by unrolling the wire on the poles in a natural manner. The system will record the covered trajectories by using a camera network composed by both 2D and 3D cameras. The learned trajectories will be translated into robot motions using a custom inverse kinematic engine. Moreover, the camera network system will be able to monitor the workspace by detecting and tracking humans, as a first step of a more advanced safely controlled environment.

  • Phase 2 (showcase): The showcase round will be focused on a specific category of motors. The learning algorithm will be targeted for winding a pole, to be mounted afterwards on a motor stator. The system will take into account the pole dimensions (height, width, and depth), the number of turns in the coil, the wire thickness and tension. These characteristics will enhance the information acquired during the free-style to compute the trajectory to be covered by the robot TCP. In fact, the considered features will be used to plan the path to wind poles never seen before. Of course, it will be still possible to refine the computed trajectory by teaching a better route, through human demonstration. A set of basic quality inspection protocols, based on turns count and wire tension, will be introduced in order to guarantee a high standard of the process. This feature enables the robot to keep constant as much as possible the wire tension, to obtain good properties of the winded coil.

  • Phase 3 (pilot): After the first two stages, the robot will be able to replicate the operator motions (winding action). When the operation ends, because of the lack of operator precision, a significant number of coils could be overlapped. Moreover, if the space between two stator teeth is narrow, the robot could face collisions problems with them. In this phase, these issues will be addressed by pursuing two objectives. The first is to exploit a safer winding action avoiding teeth collisions. The second is to reach a higher quality of coils placement around the teeth in order to increase the performance of the electrical machine (a good positioning means that all the coils are uniformly distributed over the teeth).

FLEXICOIL will ultimately produce the following scientific and technical benefits:

  • Reduction of setup times and cost of the winding machine.
  • Increase of product performances and quality.
  • Reduction of environmental impact of the production process.
  • Winding operations can be easily parallelized.

Partner of

ICT Innovation for Manufacturing SMEs



C1 Platform Introduction Day @ IPA

2015-06-29 till 2015-06-30

On Monday, the 29th, and Tuesday the 30th of June, the EuRoC members of Challenge 1 met at Fraunhofer IPA for the C1 Platform Introduction Day in Stuttgart, Germany. more

Category: eurocNews

Displaying results 5 to 5 out of 19


February 2018