Abstract – The E-brake project belongs to the Clean Sky 2 European Programme (CS2), which is a public-private partnership between the European Commission and the European aeronautics industry that coordinates and funds research activities to deliver significantly quieter and more environmentally friendly aircraft.
Introduction
E-brake is a Systems – Integrated Technology Demonstrators (SYS ITD) project, whose objective is the realization of a system to be integrated into the Piaggio Aerospace SAT (Small Aircraft Transport) technology demonstrator. The project consists in designing, producing, testing and certifying an innovative electro-mechanical brake system for SAT (Small Aircraf Transport).
Main objective
A working prototype which is fully compliant to the project requirements, comprising all the components and required documentation for the certification. The prototype of the new electro-mechanical brake system will be also integrated into the landing gear of an existing vehicle (Piaggio Aerospace P.180 EVO).
UNIPG Contributions
Coordination, Communication, Dissemination and Exploitation
Participation to the general coordination meetings and to technical meetings with partners. Participation to specific technical meetings with UAS and with Piaggio.
Technologies identification and preliminary design
The main purpose of an anti-skid control system is to maximize the braking torque transmitted to the wheel without causing it to lock. In view of such a goal, the first activity dealt with the identification of the main factors that have to be taken into account when designing the system:
- Vehicle dynamics: analysis, mathematical modelling of the physical quantities describing the motion of the vehicle on the rolling surface.
- The interaction between the vehicle and the sliding surface: mathematical models describing the interaction between the braking system (and consequently the vehicle) and the surface to describe the complete process of braking. According to the literature, those models take into account surface variation, while neglect fast tire dynamics, such as local deformations.
Detailed Design of the Actuation System and Integration
The proposed optimal-slip estimation strategy is based on a multilayer neural network (Multi Layer Perceptron-MLP). The use of a sequence of pairs allows to embed in the network the history of the braking, hence allowing to discriminate among the various possible friction curve. The MLP is trained in a supervised manner, by using a synthetic data set, that has been generated using the Burckhardt model, by sampling in two different manners the so called “Friction Cube”: a space obtained from the reference surfaces conditions, i.e. Dry, Wet, Cobblestone, Snow. The following Figure shows the selected representative points on the Friction cube as long as the corresponding (𝜆, 𝜇) Burckhardt curves.
The performance of the proposed estimation scheme is compared against a different one, from the literature [De Castro et al. (2011); de Castro et al. (2012)], which rely on the RLS strategy. The RMSE between the ground truth and the estimated optimal slip during the entire braking operation is computed to provide a quantitative evaluation. The experiments simulate the landing of an aircraft over an unknown surface whose conditions change during the braking operation. Dry, Wet, and Snow, according to the Burckhardt model, have been considered. Those surfaces where not used during the training phase of the MLP. Computational cost of the proposed MLP has been also discussed. The experiments have been performed in two different configurations:
- Open Loop -> Feedforward control
- Closed Loop
- Sliding mode controller.
- PID regulator.
