Recent advances in theory, algorithms, and computational power make it possible to solve complex, optimal control problems both for off-line and on-line industrial applications. This paper starts by reviewing the technical details of the solution methods pertaining to three general categories: dynamic programming, indirect methods, and direct methods. With the aid of a demonstration example, the advantages and disadvantages of each method are discussed, along with a brief review of available software. The main result that emerges is the indirect method being numerically competitive with the performance of direct ones based on non-linear programming solvers and interior point algorithms. The second part of the paper introduces an indirect method based on the Pontryagin Minimum Principle (PMP). It also presents a detailed procedure and software tools (named PINS) to formulate the problem, automatically generate the C++ code, and eventually obtain a numerical solution for several optimal control problems of practical relevance. The application of PMP relates to the analytical derivation of necessary conditions for optimality. This aspect-often regarded in the literature as a drawback-is here exploited to build a robust yet efficient numerical method that formally eliminates the controls from the resulting Boundary Value Problem, thus gaining robustness and a high convergence rate. The elimination of the control is obtained either via their explicit formulation function of state and Lagrange multiplierswhen possible-or via an iterative numerical solution. The paper closes presenting a minimum time manoeuvre of a car using a fairly complex vehicle model which also includes tyre saturation.

No abstract

This position paper introduces the concept of artificial "co-drivers" as an enabling technology for future intelligent transportation systems. In Sections I and II, the design principles of co-drivers are introduced and framed within general human-robot interactions. Several contributing theories and technologies are reviewed, specifically those relating to relevant cognitive architectures, human-like sensory-motor strategies, and the emulation theory of cognition. In Sections III and IV, we present the co-driver developed for the EU project interactIVe as an example instantiation of this notion, demonstrating how it conforms to the given guidelines. We also present substantive experimental results and clarify the limitations and performance of the current implementation. In Sections IV and V, we analyze the impact of the co-driver technology. In particular, we identify a range of application fields, showing how it constitutes a universal enabling technology for both smart vehicles and cooperative systems, and naturally sets out a program for future research.

No abstract

No abstract

No abstract

This article illustrates a novel Curve Warning system for motorcycles which has been developed in the SAFERIDER project (www.saferider-eu.org) of the 7th EU FP, among other Advanced Rider Assistance Systems. The Curve Warning function (CW) described here follows a holistic approach, which combines road geometry, motorcycle dynamics, rider input and riding styles. The warning strategy is based on the correction of longitudinal dynamics derived from a previewed ideal manoeuvre (reference manoeuvre) continuously computed from the actual state of the vehicle. Under normal driving conditions the reference manoeuvre matches the rider's and no correction is needed and no warning is given. But if large differences between actual and ideal accelerations are found the rider is warned to decelerate or brake. As soon as the correct value of deceleration is achieved the warning disappears, improving system acceptability. Warnings are given to the rider via an HMI, which uses a haptic accelerator throttle, a vibrating glove and helmet, and a visual display.

First and second ventilatory thresholds (VT 1 and VT 2 ) represent the boundaries of the moderate-heavy and heavy-severe exercise intensity. Currently, VTs are primarily detected visually from cardiopulmonary exercise test (CPET) data, beginning with an initial data screening followed by data processing and statistical analysis. Automated VT detection is a challenging task owing to the high signal to noise ratio typical of CPET data. Recurrent neural networks describe a machine learning form of Artificial Intelligence that can be used to uncover complex non-linear relationships between input and output variables. Here we proposed detection of VTs using a single neural network classifier, trained with a database of 228 laboratory CPET data. We tested the neural network performance against the judgement of 7 couples of board-certified exercise-physiologists on 25 CPET tests. The neural network achieved expert-level performances across the tasks (mean absolute error was 9.5% (r = 0.79) and 4.2% (r = 0.94) for VT 1 and VT 2 , respectively). Estimation errors are compatible with the typical error of the current gold standard visual methodology. The neural network demonstrated VT detecting and exercise intensity level classifying at a high competence level. Neural networks could potentially be embedded in CPET hardware/software to extend the reach of exercise physiologists beyond their laboratories.