Duccio Mugnaini is a PhD student in Computer Science and Computational Mathematics at Università degli Studi dell’Insubria, Italy. He received his master degree in Computer Science at Università degli Studi di Firenze. He has experience in the study and development of algorithms for motion design with particular focus to algebraic-geometric structure in Computer Aided Geometric Design.
The determination of a collision free path between start and goal positions through a set of obstacles is of fundamental importance for several applications, e.g. robotics, game design, automatic surgery and Computer Numerical Control (CNC). We present an integrated smooth path planner strategy that properly combines graph search algorithms and obstacle avoidance techniques with spline interpolation schemes. By focusing on planar scenarios, if only an image of the scenario is available, a proper representation of the environment can be obtained with image segmentation methods. By exploiting established graph search algorithms, we show that spline interpolation algorithms with tension parameters may subsequently provide an effective solution for the design of collision free curvature continuous paths. In particular, in order to define an obstacle avoiding curvilinear smooth path, we rely on a spline extension of polynomial Pythagorean Hodograph (PH) curves, whose first derivative components satisfy the Pythagorean condition in the real polynomial ring. Thanks to this intrinsic feature, PH curves and their spline extensions offer several advantages with respect to the standard representations adopted for computer aided design, such as polynomial arc-length functions, rational offset curves and real-time highly accurate and efficient CNC interpolators. Our path planner strategy is composed by two stages. In the first stage, an admissible piecewise linear path joining the start and goal positions is obtained by relying on obstacle avoidance techniques developed for selecting polylines with small angles within a set of collision-free solutions. In the second stage, the final smooth path is obtained by defining a suitable PH quintic spline curve interpolating the vertices of this polyline. Obstacle avoidance is ensured with an automatic choice of the tension parameters available in the scheme. A suitable selection of these shape parameters allows us to tighten the curve as much as necessary to the associated polyline. This is a joint work with Carlotta Giannelli and Alessandra Sestini.
Javier Luiso is an Electronic Engineer from Universidad de Buenos Aires, Argentina. He works for CITEDEF (Scientific and Technical Research for Defense Institute) as project manager for the Computer Graphics and Visualization Division. He has work in the design and development of training simulators for army forces. In particular the simulator for advanced observer and air defense artillery. As an engineer he is part of GPSIC, Research laboratory at the Electronics Dept. of FIUBA, with focus on signal processing, system identification and automatic control.
In time of peace, training is the most important activity for the military forces. The main goal is to allow their staff to be instructed by performing their specific activities in conditions as close as possible to the reality. Computer-graphics-based military training simulators fulfill this essential requirement as one of the most efficient ways. Argentine Army was interested in a simulator to train a complete section of the air defense artillery. A section is composed of 4 gunners that work in a coordinated way. Therefore, the simulator we built allows the simultaneous training of the section. This means that all gunners are immersed in the same virtual environment so as to be trained to avoid the attack of enemy aircrafts in a coordinated way. We describe the solution developed for the simulation of enemy aircraft's that was part of the design and development of the simulator. The behaviors of these aircraft's, the way they move and attack, were implemented with goal driven agents combined with steering behaviors.