- Fusion of the data coming from dead-reckoning and GPS sensors.
This project aims at the integration of the absolute localization information produced by a few GPS sensors on-board an outdoor vehicle, with the dead-reckoning incremental localization data produced by a set of inertial measurement units also onboard the vehicle, and with classical wheel odometry. In this work a software framework (ROAMFREE), developed for sensor fusion tasks, will be used. Suggested skills: Basic notions of Robotics. Assigned to: Scoca Luigi.
- Refactoring of the navigation system of our autonomous car.
This project aims at refactoring and improving the existing 2D navigation system of our autonomous driving vehicle. The work is intended to complement the existing algorithms and code base with new parts in order to improve its operation, its safety, its reliability, and the local motion planning module; the overall objective is to improve the interaction with mobile obstacles. Moreover, an extensive experimental activity is planned in order to consolidate the 2D demonstration infrastructure both in the area around the U5-U9 buildings as well as the area at and nearby the Villa Forno (Cinisello Balsamo) premises. Suggested skills: Basic notions of Robotics. Assigned to: Virgolino Dalila and Zanotti Marco.
- Indoor localization with iBeacons, integrating IMU dead-reckoning, for open-shelf libraries based on Dewey classification. Jointly advised with Daniela Micucci of Software Architecture Laboratory
In this project the student will review the market of iBeacon compatible devices, will select brand and model, and, after their delivery, will install them in our indoor localization demonstration area. The student will then develop an indoor localization application, based on the iBeacons, to drive a user in a specified position of the area. Additionally, in case the student will be pursuing a Laurea Magistrale thesis, she/he will develop the integration of the the iBeacon localization with the IMU-based dead-reckoning, by means of a Bayesian Filtering approach. Suggested skills: Basic notions of Robotics for the first part, introduction to Bayesian filtering for the second part. First part assigned to: Sgura Marco.
- Road layout - static objects (crossings, buildings, etc.).
In this project the student will develop the detector systems for determining the presence of some static parts of the scene observed by a forward facing cameras mounted on a urban road vehicle. These developments will be integrated in our Road Layout framework. In particular, she/he will focus on crossings and buildings. Assigned to: Cattaneo Sergio.
- Road layout - dynamic objects (vehicles, pedestrians, cyclists, etc.).
In this project the student will develop, train, and integrate in our Road Layout framework, detectors for some of the dynamic objects that can be observed by forward facing cameras mounted on a urban road vehicle. Assigned to: Meroni Valentina.
- Motion planning in dynamic environments for an autonomous car.
This project aims at the analysis of the existing state of the art in terms of literature and open source software available for motion planning for an autonomous car in conditions where the scene is inhabited by moving objects like road vehicles, pedestrians, etc. We need to take into account, in the planning, the estimated dynamic of the detected objects. Update (oct14): part of the work as been developed already, along the lines of known literature approach. Suggested skills: Basic notions of Robotics. Assigned to: Lysytsya Pavlo.
- GT Mapping - Integration of a mobile robotic platform, a 3D laser scanner, and 3D mapping software.
In this project the student is required to integrate into a working system the mobile robotic platform Robuter, the 3D laser scanner based on a Sick LMS291, the 3D incremental mapping software provided by the lab, and the global relaxation mapping software also provided by the lab. The final system will allow a human to tele-operate, via a wired joystick, the robotic platform within buildings and public outdoor environments, in order to acquire data from the laser scanner, and generate a global 3D map of the observed scene. Suggested skills: Basic notions of Robotics. Assigned to: Fabrizio Bianchini.
- Road layout - framework and detectors of baseline static parts of the road.
In this project the student will develop the software for our proposed Road Layout framework as well as she/he will demonstrate its effectiveness by developing, training, and integrating detectors for the horizontal road markings into the framework. Assigned to: Limongi Dario.
- Visual odometry - Pluri-ocular camera tracking / localization based on Lie algebra.
In this project the student will be introduced to the algebraic notions of Lie groups and will work on a C++ real-time implementation of a Lie algebra camera localization system under the ROS middleware. The student will also need to interface with real cameras for the image stream acquisition. Suggested skills: Basic notions of Robotics and Computer Vision, Fearlessness of mathematics. Assigned to: Dany Thach.
- GT Mapping - low level control of the mobile base.
This project re-uses the old robuter platform and has the following objectives: mechanics: cut away the two caster wheels, mount a new single (larger) caster wheel, so to have just 3 contact points, mount the rotary table with laser scanner on the mobile base; electronics: install the 2 new high-voltage h-bridges, select the micro-controller for the control unit (typically use an existing demo-board), interface the 2 old encoders with the micro-controller; programming: program the low-level control loops, tune the parameters of the low-level control loops, program the odometry system, program the interface with the high-level control unit (serial over usb); experimental activity: perform a large experimentation of the developments so to grant future users of the system the absence of major issues. Suggested skills: Basic notions of Robotics, Electronics and Mechanics. Assigned to: Alfredo de Santis.
- GT Mapping - Registration of the 3D point clouds taken for each view.
This project aims at the registration of the point clouds produced by the device (laser scanner mounted on the rotary table) in order to create a single large and accurate point cloud map. This project will make use of the Point Cloud library (PCL, http://pointclouds.org), a development under the ROS (http://www.ros.org) framework. Suggested skills: Basic notions of Robotics. Assigned to: Simone Ceriani.