Denis McKeown

Vehicle classification by acoustic signature

The aim of this research is to investigate the feasibility of developing a traffic monitoring detector for the purpose of reliable on-line vehicle classification to aid traffic management systems. The detector used was a directional microphone connected to a DAT (Digital Audio Tape) recorder. The digital signal was preprocessed by LPC (Linear Predictive Coding) parameter conversion based on autocorrelation analysis. A Time Delay Neural Network (TDNN) was chosen to classify individual travelling vehicles based on their speed-independent acoustic signature. The paper provides a description of the TDNN architecture and training algorithm, and an overview of the LPC preprocessing and feature extraction technique as applied to audio monitoring of road traffic. The performance of TDNN vehicle classification, convergence, and accuracy for the training patterns are fully illustrated. To establish the viability of this classification approach, initially, recordings were carried out on a strip of airfield for four types of vehicles under controlled conditions. A TDNN network was successfully trained with 100% accuracy in classification for the training patterns, as well as the test patterns. The net was also robust to changes in the starting position of the acoustic waveforms with 86% accuracy for the same test data set. In the second phase of the experiment, roadside recordings were made at a two-way urban road site in the city of Leeds with no control over the environmental parameters such as background noise, interference from other travelling vehicles, or the speed of the recorded vehicle. A second TDNN network was also successfully trained with 96% accuracy for the training patterns and 84% accuracy for the test patterns.

A FIELD TRIAL OF ACOUSTIC SIGNATURE ANALYSIS FOR VEHICLE CLASSIFICATION

The aim of this research is to investigate the feasibility of developing a traffic monitoring detector for the purpose of reliable on-line vehicle classification to aid traffic management systems. The detector used was a directional microphone connected to a DAT (Digital Audio Tape) recorder. The digital signal was pre-processed by LPC (Linear Predictive Coding) parameter conversion based on autocorrelation analysis. A Time Delay Neural Network (TDNN) was chosen to classify individual travelling vehicles based on their speed-independent acoustic signature. Locations for data acquisition included roadside recordings at a number of two-way urban road sites in the city of Leeds with no control over the environmental parameters such as background noise, interference from other travelling vehicles or the speed of the recorded vehicles. The results and performance analysis of TDNN vehicle classification, the convergence for training patterns and accuracy of test patterns are fully illustrated. The paper also provides a description of the TDNN architecture and training algorithm, and an overview of the LPC pre-processing and feature extraction technique as applied to audio monitoring of road traffic. In the final phase of the experiment, the four broad categorisations of vehicles for training the network consisted of: buses or lorries; small or large saloons; various types of motorcycles; and light goods vehicles or vans. A TDNN network was successfully trained with 94% accuracy for the training patterns and 82.4% accuracy for the test patterns.

Updating and feature overwriting in short-term memory for timbre

Previous research has demonstrated a potent, stimulus-specific form of interference in short-term auditory memory. This effect has been interpreted in terms of interitem confusion and grouping, but the present experiments suggested that interference might be afeature-specific phenomenon. Participants compared standard and comparison tones over a 10-sec interval and were required to determine whether they differed in timbre. A single interfering distractor tone was presented either 50 msec or 8 sec after the offset of the standard (Experiment 1) or 2 sec prior to its onset (Experiment 2). The distractor varied in the number of features it shared with the standard and comparison, and this proved critical, since performance on the task was greatly impaired when the distractor either consisted of novel, unshared features (Experiment 1) or contained the distinguishing feature of the comparison tone (Experiments 1 and 2). These findings were incompatible with earlier accounts of forgetting but were fully explicable by the recent timbre memory model, which associates interference in short-term auditory memory with an “updating” process and feature overwriting. These results suggest similarities with the mechanisms that underlie forgetting in verbal short-term memory.

Interference in short-term auditory memory

Although interference is a well-established forgetting function in short-term auditory memory, an adequate understanding of its underlying mechanisms and time course has yet to be attained. The present study therefore aimed to explore these issues in memory for timbre. Listeners compared standard and comparison complex tones, having distinct timbres (four components varying in frequency), over a 4.7-s retention interval and made a same–different response. This interval either was silent or included one of 15 distractor tones occurring 0 ms, 100 ms, or 1,200 ms after the standard. These distractors varied in the extent to which the frequencies of their component tones were shared with the standard. Performance in comparing the two tones was significantly impaired by distractors composed of novel frequencies, regardless of the temporal position at which the distractor occurred. These results were fully compatible with the recent timbre memory model (McKeown & Wellsted, 2009) and suggested that interference in auditory memory operates via a feature-overwriting mechanism.

Auditory memory for timbre

Psychophysical studies are reported examining how the context of recent auditory stimulation may modulate the processing of new sounds. The question posed is how recent tone stimulation may affect ongoing performance in a discrimination task. In the task, two complex sounds occurred in successive intervals. A single target component of one complex was decreased (Experiments 1 and 2) or increased (Experiments 3, 4, and 5) in intensity on half of trials: The task was simply to identify those trials. Prior to each trial, a pure tone inducer was introduced either at the same frequency as the target component or at the frequency of a different component of the complex. Consistent with a frequency-specific form of disruption, discrimination performance was impaired when the inducing tone matched the frequency of the following decrement or increment. A timbre memory model (TMM) is proposed incorporating channel-specific interference allied to inhibition of attending in the coding of sounds in the context of memory traces of recent sounds.

Intraindividual Reaction Time Variability, Falls, and Gait in Old Age: A Systematic Review

OBJECTIVES Intraindividual variability (IIV) refers to the variation in reaction time (RT) performance across a given cognitive task. As greater IIV may reflect compromise of the frontal circuitry implicated in falls and gait impairment in older adults, we conducted a systematic review of the literature relating to this issue. METHODS Searches were conducted of electronic databases that identified empirical investigations of IIV, falls, and gait in older adult samples with a mean age of 65 years or older. Data were extracted relating to IIV measures, study population, and outcomes. RESULTS Of 433 studies initially identified, 9 met inclusion criteria for IIV and falls (n = 5), and gait (n = 4). Representing a total of 2,810 older participants, all of the studies of IIV and falls showed that elevated variability was associated with increased risk of falling, and half of the studies of gait indicated greater IIV was related to gait impairment. DISCUSSION Across studies, IIV measures were consistently associated with falls in older persons and demonstrated some potential in relation to gait. IIV metrics may, therefore, have considerable potential in clinical contexts and supplement existing test batteries in the assessment of falls risk and gait impairment in older populations.

A framework for the design of ambulance sirens

Ambulance sirens are essential for assisting the safe and rapid arrival of an ambulance at the scene of an emergency. In this study, the parameters upon which sirens may be designed were examined and a framework for emergency vehicle siren design was proposed. Validity for the framework was supported through acoustic measurements and the evaluation of ambulance transit times over 240 emergency runs using two different siren systems. Modifying existing siren sounds to add high frequency content would improve vehicle penetration, detectability and sound localization cues, and mounting the siren behind the radiator grill, rather than on the light bar or under the wheel arch, would provide less unwanted noise while maintaining or improving the effective distance in front of the vehicle. Ultimately, these considerations will benefit any new attempt to design auditory warnings for the emergency services.

The role of auditory memory traces in attention to frequency

Three cued signal detection experiments demonstrated a role for auditory memory traces in frequency selectivity. The extent to which the cue predicted the signal frequency affected the size of the advantage for signals at the cue frequency over those at distant frequencies when the cue-signal gap was 10 sec but not when it was 1 sec. Detection of occasional signals presented at uncued frequencies was enhanced when they matched the frequency of cues from recent trials. With “relative” cues, which were usually followed by signals at the musical fifth above the cue frequency, performance on occasional signals at the cue frequency was enhanced relative to other unexpected frequencies. These results suggest that, regardless of the listener’s expectations and intentions, the detectability of a signal is enhanced if its frequency matches an existing memory trace. One form of voluntary attention to frequency may involve maintaining traces that would otherwise slowly decay.

Active versus passive maintenance of visual nonverbal memory

Forgetting over the short term has challenged researchers for more than a century, largely because of the difficulty of controlling what goes on within the memory retention interval. But the “recent-negative-probe” procedure offers a valuable paradigm, by examining the influences of (presumably) unattended memoranda from prior trials. Here we used a recent-probe task to investigate forgetting for visual nonverbal short-term memory. The target stimuli (two visually presented abstract shapes) on a trial were followed after a retention interval by a probe, and participants indicated whether the probe matched one of the target items. Proactive interference, and hence memory for old trial probes, was observed, whereby participants were slowed in rejecting a nonmatching probe on the current trial that nevertheless matched a target item on the previous trial (a recent-negative probe). The attraction of the paradigm is that, by uncovering proactive influences of past-trial probe stimuli, it can be argued that active maintenance in memory of those probes is unlikely. In two experiments, we recorded such proactive interference of prior-trial items over a range of interstimulus (ISI) and intertrial (ITI) intervals (between 1 and 6 s, respectively). Consistent with a proposed two-process memory conception (the active–passive memory model, or APM), actively maintained memories on current trials decayed, but passively “maintained,” or unattended, visual memories of stimuli on past trials did not.

Comparisons of Complex Sounds across Extended Retention Intervals Survives Reading Aloud

A simple experimental arrangement is designed to foil verbal rehearsal during an extended (from 5 to 30 s) retention interval across which participants attempt to discriminate two periodic complex sounds. Sounds have an abstract timbre that does not lend itself to verbal labeling, they differ across trials so that no ‘standard’ comparison stimulus is built up by the participants, and the spectral change to be discriminated is very slight and therefore does not shift the stimulus into a new verbal category. And, crucially, in one experimental condition, participants read aloud during most of the retention interval. Despite these precautions, performance is robust across the extended retention interval. The inference is that one form of auditory memory does not require verbal rehearsal. Nevertheless, modest forgetting occurred. Whatever form memory takes in this situation, it is not totally secure from disruption.