Tizar Rizano

Scalable Offline Optimization of Industrial Wireless Sensor Networks

Sensor networks are increasingly used to control and monitor industrial and manufacturing processes. In this paper, we consider the problem of optimizing a cost function for wireless sensor networks of this kind under energy consumption constraints. We focus, in particular, on the problem of coverage optimization through scheduling. Following existing approaches, we use a mixed integer linear program formulation. We show how to use partitioning techniques to decompose the problem into separate subproblems, solved individually, overcoming the exponential complexity typical of integer linear programming, while minimizing the loss in optimality. In addition, we evaluate the achieved degree of optimality by computing relatively tight bounds with respect to the optimal solution. Finally, we employ simple but effective heuristics to further improve our solution. The results show that our procedure is very efficient and scalable, and is able to find solutions that are very close to optimal. These characteristics make our approach a perfect fit for large and fixed deployments of wireless sensors, typical in factory automation and industrial applications. To show the generality of the approach, we apply our methodology to three different models of varying complexity.

Design and performance analysis of an indoor position tracking technique for smart rollators

This paper presents a position tracking technique based on multisensor data fusion for rollators helping elderly people to move safely in large indoor spaces such as public buildings, shopping malls or airports. The proposed technique has been developed within the FP7 project DALi, and relies on an extended Kalman filter processing data from dead-reckoning sensors (i.e. encoders and gyroscopes), a short-range radio frequency identification (RFID) system and a front Kinect camera. As known, position tracking based on dead-reckoning sensors only is intrinsically affected by growing uncertainty. In order to keep such uncertainty within wanted boundaries, the position values are occasionally updated using a coarse-grained grid of low-cost passive RFID tags with known coordinates in a given map-based reference frame. Unfortunately, RFID tag detection does not provide any information about the orientation of the rollator. Therefore, a front camera detecting some markers on the walls is used to adjust direction. Of course, the data rate from both the RFID reader and the camera is not constant, as it depends on the actual users trajectory and on the distance between pairs of RFID tags and pairs of markers. Therefore, the average distance between tags and markers should be properly set to achieve a good trade-off between overall deployment costs and accuracy. In the paper, the results of a simulation-based performance analysis are reported in view of implementing the proposed localization and tracking technique in a real environment.

Global path planning for competitive robotic cars

In this paper, we consider the optimal motion planning problem for an autonomous race car. A competitive autonomous car must acquire environmental and opponent information to compute, in real time, the minimum time collision free path and the low level control to track the chosen path. To cope with those requirements, we first solve the problem for a car running in isolation considering the optimal sequence of manoeuvres to approach bends and straight stretches of track. We then propose a discrete abstraction to derive a problem of graph optimisation that has a very efficient, albeit suboptimal, solution. In this context, an overtake manoeuvre against a slower car will be obtained excluding from the path the arcs that could potentially generate a collision. Finally, a control algorithm is used to ensure that the car always remains close to the planned path.

Vision-Based Robust Path Reconstruction for Robot Control

Many applications in mobile robotics require the localization of the robot with respect to the boundary of a path (e.g., a lane marker) and the reconstruction of the path in front of the robot. The former can be used as a basis for a reactive control scheme that drives the robot along the specified path and the latter is used to plan its future motion. Our solution to these problems relies on a camera mounted on the chassis and pointing headway. The relevant elements of the image grabbed from the camera (i.e., the lines delimiting the path) are robustly extracted and projected into the field-of-view of a virtual camera looking over the scene from above. This way, we obtain a top plan view that allows us to reconstruct position and bearing of the vehicle irrespective of mechanical vibrations or imperfect plane of motion. What is more, by pushing forward the virtual camera, we are able to reconstruct the path in front of the robot for some distance ahead. In this paper, we describe the main ideas underlying the approach and its implementation. The accuracy of the technique and the computational workload is evaluated through a large set of experiments.

A robotic vehicle testbench for the application of MBD-MDE development technologies

Models are used in control domains for early validation of system properties, using simulation or formal verification, and for the automatic generation of a software implementation. We propose an approach in which a functional model of the controls is matched to a model of the execution platform through an intermediate mapping model, that represents the software tasks and communication messages. The functional model is (partly) developed in Simulink and code is generated for each subsystem. Next, an abstract view of the functional model is imported in SysML. Using SysML, a model of the execution platform is created, and an implementation of the subsystems as a set of tasks and messages is defined and evaluated. The M2T Acceleo tool processes the mapping model and generates the Orocos-compliant task code executing the C/C++ functions generated from Simulink, and the inter-task communication. This paper outlines the proposed flow and provides the description of a robotic car testbench used to show the application of the methodology. The testbench has enough functional complexity and a distributed implementation to justify the creation of architecture models, while requiring a moderate cost and effort for its construction by the interested researchers.

High speed robotics with low cost hardware

High performance robotics is traditionally considered as an application area reserved to university laboratories and to the research centres of a limited group of company, which can afford high investments in equipments and system engineering. We make the point that this is not necessarily true if an adequate design is used to compensate for the lack of sophisticated sensors. To prove the validity of this idea we propose a concrete case study: driving a car-like vehicle at a high speed with a cost of the hardware below 500 Euros.

A Smart Walking Assistant for Safe Navigation in Complex Indoor Environments

Large and crowded public places can easily disorientate elderly people. The EU FP7 project Devices for Assisted Living (DALi) aims at developing a robotic wheeled walker able to assist people with moderate cognitive problems to navigate in complex indoor environments where other people, obstacles and multiple points of interest may confuse or intimidate the users. The walking assistant, called c-Walker, is designed to monitor the space around the user, to detect possible hazards and to plan the best route towards a given point of interest. In this chapter, an overview of the system and some of its most important functions are described.

Model-based design of embedded control software for hybrid vehicles

In the last decades, model based methodologies have become the mainstay of research on embedded systems development. The availability of mature computer aided tools and of well-settled industrial practices has promoted the adoption of these methodologies in large companies, which are able to amortize the cost on a large volume of products. On the contrary, the cost of software licenses and of staff training often discourages their application in small and medium enterprises. In this paper, we present a model based methodology entirely based on the adoption of open source software tools. We have applied this methodology to a real case study provided by our industrial partner proving its effectiveness.