Listening for Sirens: Locating and Classifying Acoustic Alarms in City Scenes

Marchegiani, Letizia; Newman, Paul

doi:10.1109/tits.2022.3158076

Cited by 22 publications

(15 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A wider perspective on acoustic traffic event detection was later offered by [51], [52], and [55]. In [55], the authors analyze five classes of audio events: several types of ambulance sirens, railroad crossing bells, tire screeches, car honks, and glass breaking.…”

Section: Alert Sound Detection and Recognitionmentioning

confidence: 99%

How Well Can Driverless Vehicles Hear? An Introduction to Auditory Perception for Autonomous and Smart Vehicles

Marchegiani

Fafoutis

2022

IEEE Intell. Transport. Syst. Mag.

Self Cite

View full text Add to dashboard Cite

From sirens to car horns, the urban environment is full of sounds that are designed to direct drivers' attention to conditions that require special care. Microphone-equipped autonomous vehicles can also use these acoustic cues to increase safety and performance. This article explores auditory perception in the context of autonomous driving and smart vehicles, in general, examining the potential of exploiting acoustic cues in driverless vehicle technology. With a journey through the literature, we discuss various applications of auditory perception in driverless vehicles, ranging from the identification and localization of external acoustic objects to leveraging ego noise for motion estimation and engine fault detection. In addition to solutions already proposed in the literature, we point out directions for further investigations, focusing, in particular, on parallel studies in the areas of acoustics and audio signal processing that demonstrate potential for improving the performance of driverless cars.

show abstract

Section: Alert Sound Detection and Recognitionmentioning

confidence: 99%

How Well Can Driverless Vehicles Hear? An Introduction to Auditory Perception for Autonomous and Smart Vehicles

Marchegiani

Fafoutis

2022

IEEE Intell. Transport. Syst. Mag.

Self Cite

View full text Add to dashboard Cite

show abstract

“…For humans, lower frequencies are perceptually much more important than higher frequencies and this can be represented in time-frequency representations. Gammatonegrams extend this biological inspiration, using filter banks modelled on the human cochlea and have been successfully used before in a robotics context [24].…”

Section: A Audio Classificationmentioning

confidence: 99%

Keep off the Grass: Permissible Driving Routes from Radar with Weak Audio Supervision

Williams

Martini

Gadd

et al. 2020

2020 IEEE 23rd International Conference on Intelligent Transportation Systems (ITSC)

Self Cite

View full text Add to dashboard Cite

Reliable outdoor deployment of mobile robots requires the robust identification of permissible driving routes in a given environment. The performance of LiDAR and vision-based perception systems deteriorates significantly if certain environmental factors are present e.g. rain, fog, darkness. Perception systems based on Frequency-Modulated Continuous Wave scanning radar maintain full performance regardless of environmental conditions and with a longer range than alternative sensors. Learning to segment a radar scan based on driveability in a fully supervised manner is not feasible as labelling each radar scan on a bin-by-bin basis is both difficult and time-consuming to do by hand. We therefore weakly supervise the training of the radarbased classifier through an audio-based classifier that is able to predict the terrain type underneath the robot. By combining odometry, GPS and the terrain labels from the audio classifier, we are able to construct a terrain labelled trajectory of the robot in the environment which is then used to label the radar scans. Using a curriculum learning procedure, we then train a radar segmentation network to generalise beyond the initial labelling and to detect all permissible driving routes in the environment.

show abstract

“…to have Ψ h = T ), we use zero padding at the time dimension. Specifically, we use a padding equal to ξ h for kernel shape of (3, 3), a padding equal to 2 • ξ h for (5, 5) kernel, 3 • ξ h for the (7, 7) kernel, and 5 • ξ h for the (11,11) kernel. We use no padding at the feature dimension for the f dil .…”

Section: Baseline System and Modelsmentioning

confidence: 99%

“…Also, activities of classes can overlap (polyphonic SED) or not (monophonic SED). SED can be employed in a wide range of applications, like wildlife monitoring and bird activity detection [5,6], home monitoring [7,8], autonomous vehicles [9,10], and surveillance [11,12].…”

Section: Introductionmentioning

confidence: 99%

Sound Event Detection with Depthwise Separable and Dilated Convolutions

Drossos¹,

Mimilakis²,

Gharib³

et al. 2020

Preprint

View full text Add to dashboard Cite

State-of-the-art sound event detection (SED) methods usually employ a series of convolutional neural networks (CNNs) to extract useful features from the input audio signal, and then recurrent neural networks (RNNs) to model longer temporal context in the extracted features. The number of the channels of the CNNs and size of the weight matrices of the RNNs have a direct effect on the total amount of parameters of the SED method, which is to a couple of millions. Additionally, the usually long sequences that are used as an input to an SED method along with the employment of an RNN, introduce implications like increased training time, difficulty at gradient flow, and impeding the parallelization of the SED method. To tackle all these problems, we propose the replacement of the CNNs with depthwise separable convolutions and the replacement of the RNNs with dilated convolutions. We compare the proposed method to a baseline convolutional neural network on a SED task, and achieve a reduction of the amount of parameters by 85% and average training time per epoch by 78%, and an increase the average frame-wise F1 score and reduction of the average error rate by 4.6% and 3.8%, respectively.

show abstract

Listening for Sirens: Locating and Classifying Acoustic Alarms in City Scenes

Cited by 22 publications

References 36 publications

How Well Can Driverless Vehicles Hear? An Introduction to Auditory Perception for Autonomous and Smart Vehicles

How Well Can Driverless Vehicles Hear? An Introduction to Auditory Perception for Autonomous and Smart Vehicles

Keep off the Grass: Permissible Driving Routes from Radar with Weak Audio Supervision

Sound Event Detection with Depthwise Separable and Dilated Convolutions

Contact Info

Product

Resources

About