The project senseSoar of the Autonomous Systems Lab at ETH Zurich aims at the design and autonomous operation of a small solar airplane. Whilst unmanned aircraft are already successfully applied in the defense sector, this project targets strictly non-military applications.
At a size of only 3 meters wingspan, the airplane is easy to deploy and may be operated close to the ground. We target high autonomy in the sense of flight time but also with respect to user interaction. Three application scenarios are given in the following that underline the potential of small autonomous endurance airplanes:
After disasters strike, such as earthquakes or floods, the situation often remains chaotic for several days because the overview is missing. It is also only after one or two days that high resolution satellite imagery is available to the responder teams. Using small unmanned aircraft acquiring imagery and building detailed maps would help clarify the situation, estimate damages, and even find survivors and dispatch responder teams effectively.
Search and Rescue in remote areas such as mountains or water is mainly performed by manned helicopters. However, the search part suffers from dependency on weather and visibility requirements. Also, using large swarms of helicopters for time-efficient searching is not practical due to lack of resources. The deployment of cheap small unmanned endurance aircraft has the potential of filling these gaps and allowing efficient searching also in difficult environmental conditions. Certainly, the use of a thermal camera will enhance the performance of such a system.
Areas in danger of wild fires are currently still monitored by local observation stations. Full coverage and fast localization of fire outbreaks is difficult to achieve. Therefore, constant monitoring by small solar airplanes in sustained flight would offer a new quality of monitoring and fire localization at minimal maintenance.
The Bugaboo Wildfire, 2007
Based on conceptual design considerations, a Solar Airplane prototype featuring the necessary sensors, communication capabilities and processing power is developed.
Apart from implementation of a state-of-the-art low-level controller, the focus is given to robustness in the sense of safe, failure-tolerant operation not requiring any special skills from the operator. This necessarily implies building up a terrain map on-board the airplane and also detecting other aircraft in order to safely avoid collisions with either of them.
Also, different cases of failure need to be covered: on the one hand, the controller needs to be able to detect actuator malfunctions and react accordingly both at the low level as well as at the high level with adapting the mission and regenerating paths. On the other hand, the navigation must not completely fail in the case of sensor malfunction, in particular GPS loss. Therefore, an odometry system is suggested based on vision, IMU and pressure sensors that integrates well with the GPS supported localization and mapping framework.
At the highest level, a mission task manager will need to be designed, detecting the different issues and suggesting solutions.
As an overall contribution, a guideline of how to design Small Unmanned Solar Endurance Aircraft for different mission requirements both at the conceptual level as well as at the details level is expected.
On the road of integrating Unmanned Aircraft in civil airspace, this project will attempt to offer solutions using and extending state-of-the art machine vision, sensor fusion, low and high level control making unmanned aircraft more intelligent and their operation more robust and fail-safe. The following three topics are seen as the most important core technology:
Small and slow flying endurance UAS require lightweight actuators that have a comparably high failure rate. It is thus crucial to offer solutions of detecting and compensating failures. Since system modeling is a prerequisite for this technology, power-efficient control strategy development by making use of the redundancies will be going hand-in-hand. The resulting fail-safe and efficient controller compensates one of the major deficiencies of small and slow UAS, but it will be equally applicable to UAS of any size and weight increasing their robustness and safety.
Autonomous UAS operation implies the capability of intelligently detecting terrain and intruder aircraft in order to also automatically react. Research in this area is ongoing, but lightweight, low-power and low-cost options do still not exist. Using vision as the primary means of range sensing is the direct analogy to a human pilot in Visual Flight Conditions (VMC). The fusion with IMU, GPS and pressure information is expected to provide a powerful and robust solution at minimum weight and power consumption which is platform independent. While collision warning systems such as TCAS or Flarm are well established, they rely on all intruder aircraft to have them installed. Thus, the non-cooperative sense-and-avoid technology to be developed is regarded very useful also for future UAS and when it comes to their integration with civil airspace. Last but not least, it is also imaginable to be used as pilot assistance in manned aviation allowing a further improvement of flight safety.
Easy operation of highly autonomous UAS was set a primary goal of this project and is a hot research topic. The last but very important level of autonomy is provided by automatic landing. While GPS or laser guided landing to a fixed installation has been a solved problem, the small UAS should be given the capability of landing without such further installations also at remote locations in emergency scenarios. Small UAS would experience a dramatic gain in interest upon the successful demonstration of such technology – rendering them more flexible in operation, more failure-tolerant and safer compared to existing systems.
| Project Lead
Stefan Leutenegger, ETH Zurich
Autonomous Systems Lab
CLA E 16.2
Phone: +41 44 632 24 28
Thomas Baumgartner, ETH Zurich
Autonomous Systems Lab
CLA E 20
Phone: +41 44 632 57 76
| Control System
Konrad Rudin , ETH Zurich
Autonomous Systems Lab
CLA E 16.2
Phone: +41 44 632 89 49
| Aerodynamics and Lightweight Structure
Phone.: +49 531 516 89 0
Fax: +49 531 516 89 29
|Conceptual and Preliminary Design||2009|
|Detail Design, Manufacturing and Assembly||2010-2011|
|First Flight||Summer 2011|
|Vision Support and Mapping||2011-2012|
|Navigation and High-Level Control||2011-2012|