Raphaël Labayrade

Cooperative Fusion for Multi-Obstacles Detection With Use of Stereovision and Laser Scanner

We propose a new cooperative fusion approach between stereovision and laser scanner in order to take advantage of the best features and cope with the drawbacks of these two sensors to perform robust, accurate and real time-detection of multi-obstacles in the automotive context. The proposed system is able to estimate the position and the height, width and depth of generic obstacles at video frame rate (25 frames per second). The vehicle pitch, estimated by stereovision, is used to filter laser scanner raw data. Objects out of the road are removed using road lane information computed by stereovision. Various fusion schemes are proposed and one is experimented. Results of experiments in real driving situations (multi-pedestrians and multi-vehicles detection) are presented and stress the benefits of our approach.

Real-time disparity contrast combination for onboard estimation of the visibility distance

An atmospheric visibility measurement system capable of quantifying the most common operating range of onboard exteroceptive sensors is a key parameter in the creation of driving assistance systems. This information is then utilized to adapt sensor operations and processing or to alert the driver that his onboard assistance system is momentarily inoperative. Moreover, a system capable of either detecting the presence of fog or estimating visibility distances constitutes in itself a driving assistance. In this paper, the authors present a technique to estimate the mobilized visibility distance through a use of onboard charge-coupled device cameras. The latter represents the distance to the most distant object on the road surface having a contrast above 5%. This definition is very close to the definition of the meteorological visibility distance proposed by the International Commission on Illumination. The method combines the computations of local contrasts above 5% and of a depth map of the vehicle environment using stereovision within 60 ms on a current-day computer. In this paper, both methods are described separately. Then, their combination is detailed. The method is operative night and day in every kind of meteorological condition and is evaluated; thanks to video sequences under sunny weather and foggy weather.

A collision mitigation system using laser scanner and stereovision fusion and its assessment

Road traffic incidents analysis has shown that 52% of them are caused by a collision between two vehicles or between a vehicle and an obstacle. In this paper, a collision mitigation system is proposed and evaluated towards various typical road situations. The aim of the system is to decrease the kinetic energy of the collision through automatic emergency braking that occurs 1 second before the collision. This emergency braking is triggered by an alarm coming from a decision unit taking into consideration the results of a generic obstacles detection system -based on fusion between stereovision and laser scanner- and a warning area in front of the vehicle. The different sub-systems are presented. Various typical dangerous road situations are then introduced. The behavior of the collision mitigation system towards these situations is presented through real tests carried out in the context of the ARCOS French project. These experiments show the reliability, the genericity and the efficiency of the system. In particular, the false alarm rate is low, the detection rate is high and the system proves to be reactive.

Long Range Obstacle Detection Using Laser Scanner and Stereovision

To be exploited for driving assistance purpose, a road obstacle detection system must have a good detection rate and an extremely low false detection rate. Moreover, the field of possible applications depends on the detection range of the system. With these ideas in mind, we propose in this paper a long range generic road obstacle detection system based on fusion between stereovision and laser scanner. The obstacles are detected and tracked by the laser sensor. Afterwards, stereovision is used to confirm the detections. An overview of the whole method is given. Then the confirmation process is detailed: three algorithms are proposed and compared on real road situations

A multi-model lane detector that handles road singularities

Road detection through the use of an onboard camera is an essential task required for the development of advanced driving assistance systems. Although many solutions have been proposed in literature, most existing algorithms are not designed to handle road singularities such as freeway exits, double painted lines, zebra road markings, etc. Such conditions can be confusing for vision-based algorithms although they are often met in real driving situations. In order to tackle this issue, we propose the use of various instances of a road model at the same time. We first describe the generic road model used and then explain how the different instances of this model are updated. The extracted models can be used either to detect the correct vehicle lane or to obtain a more complete description of the road configuration. Experiments are presented to demonstrate the validity of the approach. An application to multi-lane detection is also presented

On the design of a single lane-markings detectors regardless the on-board camera's position

In this article, we present an algorithm for lane marking features extraction and robust shape estimation of lane markings. The algorithm uses a new lane-marking features extractor followed by the robust fitting algorithm described by Tarel et. al. (2002) to estimate the lane-markings shape as a single analytical curve. The lane-marking features extractor is based on the fact that lane-markings widths are in a small range of possible values, on a road. This implies a geometric constrain on the observed lane-markings widths from a camera on-board a vehicle. The lane-marking features extractor uses this property to select pairs of edge points corresponding with a high probability to a section of lane-marking. This features extractor is especially designed to be robust to lighting conditions as shown by few experiments. The extracted features are then grouped to estimate the parameters of the analytical curve model of these lane-markings. With the proposed approach, the obtained detector is robust to different kinds of noise and perturbations, allowing us to use it with a camera in many positions. Finally, to illustrate this property, we briefly describe two applications of the proposed detector: accurate vehicle location on the road and estimation of the time to line crossing.

Experimental Assessment of the RESCUE Collision-Mitigation System

Road-traffic-incident analysis has shown that 52% of incidents are caused by a collision between two vehicles or between a vehicle and an obstacle. In this paper, the REduce Speed of Collision Under Emergency (RESCUE) collision-mitigation system (version 1.0) is presented and evaluated toward various typical road situations. The aim of the RESCUE system is to decrease the kinetic energy dissipated during a collision through automatic emergency braking that occurs 1 s before the collision. This emergency braking is triggered by an alarm coming from a decision unit taking into consideration the results of a generic obstacle-detection system-based on fusion between stereovision and laser scanner-and a warning area in front of the vehicle. The different subsystems are presented. Then, the behavior of the RESCUE collision-mitigation system toward various typical dangerous road situations is assessed through systematic tests. These quantitative tests are completed by qualitative ones carried out on 737 km of open roads (freeways, highways, rural roads, downtown) to provide a more precise idea about the false-alarm rate. The experiments show the system is promising in terms of reliability, genericity, and efficiency

Multi-objective optimisation of lighting installations taking into account user preferences – a pilot study

A multi-objective methodology is proposed to optimise lighting solutions by taking into account not only energy objectives but also subjective ones such as visual preferences. The originality of this method lies in the integration of subjective data obtained from psycho-visual tests conducted with panels of observers potentially representative of end users, into a multi-objective optimisation framework. This method makes it possible to identify all the best tradeoffs and the direct correspondence between power demand and user visual preferences, to handle constraints, to handle inter-individual differences and to optimise preferences on various visual attributes at the same time. The methodology was developed through a pilot study of office lighting optimisation. The findings are consistent with previous work and provide additional application knowledge about the case study.

Validation of an online protocol for assessing the luminous environment

We developed a web application to investigate online-based experiments for the psychovisual assessment of luminous environments. We reproduced online a laboratory experiment conducted in a virtual environment, in which the luminous environment of an office room lit by different combinations of ambient/task lighting was assessed by observers. A control panel of 30 observers performed both experiments. Statistical analyses revealed no statistically significant differences between data collected in the laboratory and online. In addition, an online-based experiment involving 1114 observers was conducted to study the uncontrolled experimental conditions that may impact results. We show that, by increasing the panel size, bias, mainly related to the perceived contrast and brightness of the display and the brightness of the surrounding area is removed. Findings suggest that 100 observers are enough to remove bias.

Multi-objective Optimization under Uncertain Objectives: Application to Engineering Design Problem

In the process of multi-objective optimization of real-world systems, uncertainties have to be taken into account. We focus on a particular type of uncertainties, related to uncertain objective functions. In the literature, such uncertainties are considered as noise that should be eliminated to ensure convergence of the optimization process to the most accurate solutions. In this paper, we adopt a different point of view and propose a new framework to handle uncertain objective functions in a Pareto-based multi-objective optimization process: we consider that uncertain objective functions are not only biasing errors due to the optimization, but also contain useful information on the impact of uncertainties on the system to optimize. From the Probability Density Function (PDF) of random variables modeling uncertainties of objective functions, we determine the ”Uncertain Pareto Front”, defined as a ”tradeoff probability function” in objective space and a ”solution probability function” in decision space. Then, from the ”Uncertain Pareto Front”, we show how the reliable solutions, i.e. the most probable solutions, can be identified. We propose a Monte Carlo process to approximate the ”Uncertain Pareto Front”. The proposed process is illustrated through a case study of a famous engineering problem: the welded beam design problem aimed at identifying solutions featuring at the same time low cost and low deflection with respect to an uncertain Young’s modulus.