Call3 Monitoring RIVERWATCH

Public Summary Month 3/2013

Sat, 27 Apr 2013 17:50:02 GMT

This document reports the progress of the RIVERWATCH experiment from week 51 to week 59.

Significant advances were made in the period, in particular related with the acquisition which was finalised and the work carried out for the development of the UAV plataform.

The main tasks related with hardware integration during this period was the hardware integration of all the components (hardware and software) for the quadcopter. In regarding the ASV development the new hexacopter model was integrated in our simulator. Outdoor tests involving the vision based detector for the UAV were carried out. More importantly several advances in the operation centre were carried out since the last report, with some more modules developed/adapted to the system.

Public Summary Month 1/2013

Sat, 27 Apr 2013 11:13:05 GMT

This document reports the progress of the RIVERWATCH experiment from week 42 to week 50.

The acquisition process is not finished yet which impacts some of the tasks. The Airmar PB200 Weather Station and the Sonar are tested and integrated.The robot's 3D perception module's is being developed and the results in land were very promising. The next move is to test it in water surface conditions.The UAV vision based detector proceeded and improved. The work on image processing for terrain classification is showing promising results.

Public Summary Month 11/2012

Tue, 04 Dec 2012 00:42:07 GMT

This document reports the progress of the RIVERWATCH experiment from week 33 to week 41.

Progress has been achievedin most tasks. Some of the components have already arrived while others are still in the acquisition process because for instance of customs clearance. The water-land segmentation process that exploits a priori knowledge of the shoreline pose, extracted from overhead imagery was further developed and we are currently implementing a symmetry detection operator in order to exploit the mirroring effect observed in settled waters. Then, these cues will be fused for a robust extraction of the shoreline. regarding energy harvesting and management, we have specified the problem and we expect to have an instantiation of the model available for simulation during the next reporting period. During the period the USV vision based detector was further developed and the design of the UAV overhead image processing for terrain classification procee has been done and its implementation initiated. In this report we overview the overall idea with the expectation that details can be provided in the forthcoming report.

Public Summary Month 10/2012

Mon, 15 Oct 2012 12:33:18 GMT

This document reports the progress of the RIVERWATCH experiment from week 25 to week 32.

Most of the tasks have progressed smoothly.

At the time of writing this report, Uninova received the prefunding what allowed to start contacting the suppliers regarding the order of several components essential to the project.

In Task 3 we are developing a water-land segmentation process that exploits a priori knowledge of the shoreline pose, extracted from overhead imagery. To attain this goal, we have previously reported a preliminary step, namely the segmentation of the sky region from the rest of the image. Motivated by the work of Achar et al. (2011), we have developed a mechanism, during the current reporting period, for the generation of training data from the sky/non-sky segmentation process.

In task 4 the work in the simulation environment proceeded as well as the USV vision based detector.

Public Summary Month 7/2012

Wed, 03 Oct 2012 13:23:34 GMT

This document reports the progress of the RIVERWATCH experiment from week 17 to week 25.

The work progressed in all tasks.

The major results in the period are: 1) the segmentation of the sky region from the rest of the image which was concluded in task 3; 2) the USV vision based detector progressed and some results are shown; and 3) the work of the control centre in task 6 progressed and a new version was created.

Public Summary Month 5/2012

Mon, 17 Sep 2012 16:54:28 GMT

This document reports the progress of the RIVERWATCH experiment from week 8 to week 16.

All tasks progressed considerably in this reporting period.

During the reporting period all the equipment has been selected and quotation requests issued.

The development of the simulation environment was carried out. It is composed of two modules, one responsible for simulating the presence of obstacles above the surface and another responsible for simulating the surface’s dynamics and the interaction of the underwater sonar beams and riverbed.

A first version of the software architecture for the ASV navigation system was proposed in the previous report. In the current reporting period, several aspects of this architecture has been partially instantiated and tested on a simulated vehicle with double Ackerman kinematics. Validation on the extended USV simulation architecture will be carried out in the forthcoming reporting period.

A key component of the UAV navigation system is the USV vision-based detector. Without it, the UAV will not be able to properly dock on the USV. The detector has been devised and will be implemented in the forthcoming reporting period. We expect to have results soon enough to submit this work to the next IROS conference.

The control centre is almost finalised, requiring now a thorough field evaluation. Due to the impossibility of running field tests in the short-term, the effort put in this task during the reporting period was directed towards bug fixing and minor adjustments at the user interaction level.

Public Summary Month 3/2012

Mon, 17 Sep 2012 16:29:47 GMT

The project aims at enabling environmental monitoring of rivers with a robotic team composed of an autonomous surface vehicle (ASV) and a quad-copter. The major challenges to be faced in the project are related to autonomous navigation in riverine environments and complex interactions between the two robotic platforms. There is also considerable integration effort involved.

This document reports the progress of the RIVERWATCH experiment from week 1 to week 8. In this period, tasks 1, 2, 3, 4, and 6 advanced as planned in the experiment’s proposal.