Periodic Reporting for period 2 - ODESSA (Obstruction DEtection Sensor for Surveillance on Aircraft)
Reporting period: 2019-10-01 to 2020-12-31
Mid-air, near mid-air, but also on ground collisions with obstacles are considered significant cause of accident in general aviation. For this reason, the design of an affordable obstacle warning sensor, helping pilots in preventing accidents could be very important.
The use of millimetric radar, derived from automotive domain, gave different advantages. Reliability and solutions that are small, light, low consumption and low cost.
ODESSA system improves the safety of landing and ground procedures, being independent from the airport infrastructure.
Achieved key objectives:
• Build up a prototype of ODESSA Sensor, based on millimetric radar technology combined with a video camera (A and B-models ), suitable for large scale production of “on board” certified equipment
• Build up a Control Unit, which provides a smart graphic HMI, to manage the data incoming from the sensors. Image Processing algorithm (AI techniques) was developed to enable the target detection and recognition capabilities
• Perform the characterization of the prototypes both in flight (RPAS payload), and in ground, by simulating different environmental scenarios like rain, smoke, Ice spray, etc. More than 60 different tests have been carried out
• Optimization of Size, Weight and Power consumption to be compatible with the aeronautic constraints, achieving consequently one of main topic of CleanSky2: reducing emission, optimising aircraft design
The Control-Unit weight is 1.8 kg with a Power Consumption of 30 W, small footprint (230x216x102 mm). The Sensor-Unit is less than 1 kg with a Power Consumption of 18 W.
key technical results:
• Linear Range detection: up to about 150 meters. The test limited up to 138 m, due to the test field area limitation
• Target speed: The tests in flight, and at during ground test showed the capacity to detect moving targets at the relative speed of 10-50Km/h
• Elevation detection : vertical FoV 15° total, with the use of three sensors installed in in the ODESSA B-model equipment
• Azimuth detection: horizontal FoV of a single equipment is ± 40°. In order to reach the 80° azimuth, the final system application must be configured simply installing two sensors, placed side by side to cover a greater horizontal FoV
• Target recognition and identification: different targets have been used for the Target recognition and identification capability characterization: Cables (8mm, 16mm diameter), unanimated and animated objects such as car, tree, building, high voltage pylon, human, aircraft, small UAV, corner reflector (target simulator). During test case, all objects have been detected by the sensor and recognized
• Number of target: the sensor showed the capability to detect and identify more than 10 obstacles simultaneously.
• ODESSA WEB Site
• Twitter: @Odessa_project
• LinkedIn: Odessa Project
IC participated to many important public events, where ODESSA “live“ prototypes were presented; most important was the SIAE-2019, presenting a working prototype.
A remote conference @Sapienza Università di Roma was also held.
The paper “Design of a Substrate Integrated Waveguide Slots Antenna in W Band for Aircraft Radar Application” has been presented at by Prof. Dauvignac (CNRS) on behalf of ODESSA consortium at the IEEE Radar conference, Florence
The dissemination of the project has been performed by dedicated tools (project flyable, project Roll up, animated video). All the dissemination material is available on the ODESSA website
Outlook of A/B-model implementation
Odessa System is based on a flyable equipment (Sensor Unit) and a remote control equipment (Control Unit), connected thorough a dedicated data link
Two developments and test campaigns were carried out (A-model prototype and industrialized equipment, B-Model). A state of the art antenna has been explicitly designed to meet the range detection requirements.
The first prototype of the ODESSA system, including the sensor prototype and the utilities for system test, was tested by using a COTS Processing Unit installed into a Box realized by means of a 3D printer.
The second prototype of the ODESSA (B-mod) was assembled into a custom mechanical chassis, able to host up to three sensors, in order to enhance the azimuth performance. An integrated electronic Processing module was developed to implement the Control Unit.
Intensive research, simulation and development activities was carried out by CNRS on the RX Antenna, to properly operate in the frequency range of interest (76 to 77 GHz); several different layout have been tested, too.
A SW graphic interface was developed by IC, to manage the ""Raw Data"" and ""Target List"" provided by the sensors: the HMI allows the Flight Operator to Visualize Sensors Data (targets position, camera video, alarm information) and to Interact with the System.
A dedicated SW application was developed. In particular a Convolutional Neural Networks for the classification of the detected obstacles, classified in categories, was implemented: this AI algorithm represents one of the key result of the research.
In Flight Sensor operational data collection:
The partner SIRALAB provided the RPAS suitable for the project trials.
Hundreds flight and ground tests have been performed against the A-model and B-model (Flight/Ground Test Plans).
The objective was to characterize the behaviour of the sensor in presence of different types of target (Different Cables , stationary and mobile objects like cars, trees, building, high voltage pylon, human, aircraft, small UAV, corner reflector used as target simulator).
Tests have been carried out in different environmental conditions (rain , fog/smog, Icy..) in order to characterize the system behaviour in all-weather meteo situations.
The first flight (A-mod) was performed on 16/11/2019, while first flight with full ODESSA B-model on early November 2020. Collected Data available on the ODESSA portal and in the public domain Zenodo.
An important result of the flight campaign is that no interference was detected during flight trials between radar operation and both control and wireless data links.
Then the capability of install the ODESSA sensor unit on a remotely piloted platform has been demonstrated.
The use of the Control Unit for manned platforms (e.g. helicopters) will need a further ruggedization and qualification.
The present B-model is more suitable for VTOL airborne platforms (piloted or remotely piloted)."
The main outcome of the research is also that a reliable and light anti-collision system is feasible and fully applicable to VTOL platforms, including RPAS.
Use of RPAS is significantly growing, and anti-collision system will be very useful to improve safety, giving positive impact in RPAS industries.