Adaptative Robotic Algorithms for Sanding (ARAS)

In the last decade, robots in the manufacturing industry have increased their presence by handling activities out of the automated processes traditionally operated by industrial robots. From robotics cells, in which big and costly robots performed repetitive operations, to a flexible environment were robots collaborate with human workers in manufacturing activities, the robotics paradigm is changing in the manufacturing industry worldwide.

ARAS (Adaptative Robotics Algorithms for Sanding) pursues the conception and development of new approaches for robotizing manufacturing operations as sanding or polishing.

The project aligns with the current trends of the manufacturing industry of using cobots (collaborative robots) that could work in the same environment of human operators and with the capability of dynamically adapt to the different shapes of the pieces to be manufactured.

ARQUIMEA has developed software libraries that allow lower cost, faster implementation, ease of use and higher quality (repetitive and uniform) automated sanding and polishing operations in the industry 4.0.

ARAS Adaptative Robotic Algorithms for Sanding


Manufacturing operations as polishing or sanding have not been extensively automated yet. The reason is that traditional approaches used in robotics cells are based on learned trajectories, programmed by a human operator for performing repetitive tasks in which the environment and the piece under manufacturing are perfectly known and controlled during the whole process. Sanding or polishing are complex operations that require an adaptative dynamic solution, i.e. the robot shall adapt its trajectory and force dynamically during the process. This becomes more challenging if the piece under manufacturing is not perfectly known (dimensional uncertainties), if the piece is not flat (curvature is present in its surface) or if the piece bends during the operation due to its flexible properties.

ARQUIMEA pretends to solve this challenge developing algorithms able to dynamically adapt to unknown 3D surfaces or flexible materials which cannot be predefined.



Different solutions using specific hardware have recently been released to the market for curved pieces sanding and polishing. ARQUIMEA implemented a solution based on software with the aim of reducing costs of complex hardware elements and making a product compatible with collaborative robots.

The solution is based on a set of libraries adaptable to the cobot programming language that allows a fast and easy parametrization of the desired sanding or polishing process to be done. The operator just provides the desired parameters of force, velocity, trajectory characteristics using the teach pendant of the cobot. During the execution phase, ARAS automatically generates the path following user indications and controls the robot movements to adapt to the surface under processing. This is done using a force sensor installed in the robotic arm. The existing force applied at each point of the workpiece is available to the user for later quality analysis using industry 4.0 capabilities.

In addition, ARQUIMEA´s collaboration with 3M allows to provide counselling for the selection of the best world leading abrasives in the market and how they should be used in each specific application for an improved quality result matching customer needs.


The main result is a library of functions that the programmer can easily use in their robot programming environment for dealing with sanding and polishing processes of the most challenging pieces such as airplane wings, surface preparation for mechanical parts treatments, ships, wind power generators blades, automotive industry, etc…

The sanding and polishing process is quite straight forward and follows the next steps:

  1. The operator teaches a few outline points of the objective area. There is no need of teaching the curvature or internal points but just delimitation of the outline area for sanding or polishing.
  2. The algorithm automatically generates the path planning by following user preferences of force, orientation and velocity.
  3. The robot dynamically adapts to the surface curvature controlling the force during the execution.
  4. Finally, a quality control of the process is available: 3D modelling of the sanded surface and force applied in each point allowing
ARAS – Adaptative Robotic Algorithms for Sanding

Ready to take the leap?
Let us help you.