The latest developments in automobile design have allowed them to be equipped with various sensing devices. Multiple sensors such as cameras and radar systems can be simultaneously used for active safety systems in order to overcome blind spots of individual sensors. This paper proposes a novel sensing technique for catching up and tracking an approaching vehicle relying on an acoustic cue. First, it is necessary to extract a robust spatial feature from noisy acoustical observations. In this paper, the spatio-temporal gradient method is employed for the feature extraction. Then, the spatial feature is filtered out through sequential state estimation. A particle filter is employed to cope with a highly non-linear problem. Feasibility of the proposed method has been confirmed with real acoustical observations, which are obtained by microphones outside a cruising vehicle.

This paper proposes a robust sensing technique for localizing an approaching vehicle relying on an acoustic cue. A camera and a radar system are commonly used as sensors for active safety systems. However, such sensors do not work well in a blind spot such as a blind junction of a highway. On the other hands, various acoustical noises caused by a cruising vehicle arrive at the blind spot. Those acoustic cues are available for achieving robust sensing of surrounding vehicles. It is necessary to extract a robust spatial feature from noisy acoustical observations. In this paper, the spatio-temporal gradient method is employed for the feature extraction, and the spatial feature is filtered out through sequential state estimation. Feasibility of the proposed method is confirmed with real acoustical observations, which are obtained by microphones outside a cruising vehicle. The filtering process can be done in real time, because its real-time factor is 0.096 using a 2.6 GHz Intel Core i7 processor.

No abstract

Recently, new cell phone services enable taking input by speaking to a quasi-agent using highly accurate speech recognition technologies. However, there are two problems when equipping a vehicle with these technologies. First, we do not understand yet the effect on driver’s usability performance of a car navigation system equipped with spoken dialogue interface. Second, the corresponding design techniques to obtain sufficiently good navigation results has not yet been determined. In this study, we first discuss problems involving speech recognition technologies and the types of speech interface used in this study. Second, we experimentally compare in pseudo-driving conditions the effect on driver’s physiology of various speech user interfaces: the Command Input (CI) interface, the Natural Language Understanding (NLU) interface, and the NLU/Spoken Dialogue Management (SDM) interface. According to the analysis of physiological signals, parasympathetic nervous system activity index value was higher when using the NLU/SDM interface than when using the CI interface. Furthermore, both subjective assessment result and time satisfaction degree of NLU/SDM interface were higher than others. Therefore, we can conclude that the NLU/SDM-type interface has better usability performance.

In recent years, speech interfaces have been proposed for making shopping order or restaurant reservations while transmitting information only by speech. It is easy to imagine that such a system will be installed in cars soon. However, to use the speech interface that can realize such tasks while the driver is driving, the system needs to understand the driving situation of the driver and control the speech guide timing accordingly. We focused on the fact that there are scenes (SCD: Scenes of Concentrate Driving) where the driver temporarily interrupts the speech operation unconsciously to concentrate on driving operations. However, one does not yet know in which scenes the voice dialogue becomes a burden. In this study, we identified driving scenes where dialogue is a burden (SCD), clarified the relationship between SCD and driving behavior, and considered a method to estimate SCD automatically. In this research, we define SCD as a situation in which the driver temporarily suspends the speech operation and wants to perform only the driving operation, although the system can perform both the speech operation and the driving operation. Under SCD circumstances, even if the speech interface presents some guidance, the driver can not understand its contents. Also, even if the system prompts for an answer by speech input, the driver does not answer. To confirm the existence of SCD, we have created nine driving scenes that require different driving operations (lane change, overtaking, narrow road, right turn, etc.) and reproduced these scenes using a Driving Simulator (DS). We also created some tasks to prompt the driver's voice input, and present them just before the driving scene, such we could check whether the driver interrupts or not the dialogue during the driving operations needed to pass the scene. We recorded speech and collected driving signals (steering angle, throttle opening, etc.) for 15 men and women in their 20s (11 men and 4 women) with a driver's license. First, for each driving scene, we checked whether the driver could answer the immediately preceding question presented by an operator and confirms that SCD occurs in various driving scenes. In the driving scenes involving lane change, the utterance was interrupted for more than 40% of the tasks. On the other hand, in the driving scenes in which the driver steers left and right curves, while driving in the same lane as the preceding vehicle, the utterance was seldom interrupted. These results confirm the existence of driving scenes in which speech is likely to be interrupted, that is, SCD. Next, we considered a machine learning model to estimate the occurrence of SCD automatically, using only the driving signal. We compared the classification accuracy of SCD occurrence according to several analysis window lengths, from 0.5seconds to 5 seconds with a 0.5seconds increment. The results showed that a trained Support Vector Machine (SVM) could provide a classification accuracy of about to 85% for two classes (SCD or not), and 82% for three classes (SCD, semi-SCD, regular) with a window size of 2seconds.