Anagrafe della ricerca

For the next generation of farmer robots, it is critical to address open challenges towards safe and efficient deployment in highly challenging environments. Precision Agriculture tasks like crop monitoring, spraying, weeding, canopy management as well as different agronomic conditions could require working in different modalities. This is especially true on steep and rugged terrains where novel solutions need to be found for efficient and safe locomotion as well as to increase the level of autonomy for long duration applications. Within the FACE project, three main issues for electrically driven agricultural robots will be directly tackled: traction control, terrain awareness and rollover mitigation. Off-road vehicles adopt usually multiple driven axles, thus, calling for a suitable control strategy that minimizes energy consumption by preventing wheel slippage while ensuring optimal traction and steerability performance even in the presence of narrow sloped passages and steering through crop rows. In addition, enabling a farmer robot with the ability to sense the traversed terrain would make it possible to optimize its locomotion performance by implementing site-specific control and planning strategies. In contrast to existing methods that mostly rely on forward imaging sensing failing to provide any information about the mechanical properties that directly impact on vehicle mobility, in the context of FACE the ambition is to rely on proprioceptive sensors (e.g., accelerometers and torque sensors). The envisaged idea is that terrain category can be identified using wheels/tracks as tactile sensors that generate signals modulated by the vehicle-terrain interaction. Increasing robot awareness on the rollover risk and implementing safety measures is another key issue. Current efforts have been devoted to safety systems for tractors focusing on dynamic simulations and rather simple rollover indicators. However, much research has to be developed yet: episodes of impending tire contact loss have to be recognized and treated, and dynamic indicators as well as correction actions, e.g., active steering and body reconfiguration, are new areas of investigation. One important aspect of the FACE project is that it will cover both theoretical and experimental research. On the one hand, activities will be devoted to understanding the fundamental physical mechanisms underlying robot-environment interaction, and to the development of physical models and estimation methods that allow automated machines to operate efficiently and safely in complex unstructured scenarios. On the other hand, rigorous experimental analysis and demonstration will be performed leveraging on existing partners’ robots and facilities. One unique feature is the use of heterogeneous farmer robots including multi-wheeled, tracked and reconfigurable articulated prototypes, which gives the opportunity to demonstrate the scalability and general applicability of the proposed technologies.

• LIBERA UNIVERSITA' DI BOLZANO - Coordinatore
• POLITECNICO DI BARI
• POLITECNICO DI TORINO - AMMINISTRAZIONE CENTRALE