Insect-inspired vision for autonomous vehicles

Serres, Julien; Viollet, Stéphane

doi:10.1016/j.cois.2018.09.005

Cited by 14 publications

(6 citation statements)

References 57 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With a growing interest in bio-inspired technology, many recent studies have focused on the development of vision sensors inspired by insect eyes for autonomous vehicles (Serres and Viollet, 2018). A particular focus has been placed on understanding and instantiating the physiological and motion processing mechanisms that make insects such proficient flyers.…”

Section: Discussionmentioning

confidence: 99%

Mechanisms of punctuated vision in fly flight

Cellini

Salem

Mongeau

2020

Preprint

View full text Add to dashboard Cite

To guide locomotion, animals control their gaze via movements of their eyes, head, and/or body, but how the nervous system controls gaze during complex motor tasks remains elusive. Notably, eye movements are constrained by anatomical limits, which requires resetting eye position. By studying tethered, flying flies (Drosophila) in a virtual reality flight simulator, we show that ballistic head movements (saccades) reset eye position, are stereotyped and leverage elastic recoil of the neck joint, enabling mechanically assisted redirection of gaze. Head reset saccades were of proprioceptive origin and interrupted smooth movements for as little as 50 ms, enabling punctuated, near-continuous gaze stabilization. Wing saccades were modulated by head orientation, establishing a causal link between neck signals and execution of body saccades. Furthermore, we demonstrate that head movements are gated by behavioral state. We propose a control architecture for biological and bio-inspired active vision systems with limits in sensor range of motion.

show abstract

Section: Discussionmentioning

confidence: 99%

Mechanisms of punctuated vision in fly flight

Cellini

Salem

Mongeau

2020

Preprint

View full text Add to dashboard Cite

show abstract

“…The rapid and accurate identification of insects is essential whereby the prevention of economic losses and its contribution to the field of entomology [6], [7]. Also, insects inspire scientists in robots, sensors technologies, mechanical structures, aerodynamics, and intelligent systems [8]. The estimated number of species in insects is 1.5 million on the earth, but the number of named and…”

Section: Introductionmentioning

confidence: 99%

Comparison of Deep Learning Techniques for Classification of the Insects in Order Level With Mobile Software Application

ÖZDEMİR

Kunduraci

2022

IEEE Access

View full text Add to dashboard Cite

Insects are a class of the arthropod branch and the most crowded animal group in terms of species and taxonomy. Due to destruction and forest fires, some insect species could go extinct without being detected. Identifying new insects and having knowledge about insects in terms of biodiversity will contribute positively to the studies carried out, especially in entomology, agriculture, the pharmaceutical industry, medicine, robotics, and other branches. In this study, we produced a mobile-based decision support software with a deep learning model to classify and detect insects at the order level. We also presented the comparative analysis results of SSD MobileNET, YoloV4, and Faster R-CNN InceptionV3 deep learning methods and adapting processes for order-level insect classification. Our approach studies the suitability of existing models towards such an objective, and we conclude that Faster R-CNN InceptionV3 performs the best at classifying and detecting insects at the order level. In addition, we shared 25820 training and 1500 test data in the kaggle database in order to contribute studies to be carried out in this area. As a result, we believe that this research will be beneficial to entomologists, naturalists, and other researchers in related fields.

show abstract

“…To fully demonstrate the advantages of this approach we use the example of autonomous navigation as it is well studied and algorithmically understood in a variety of environments be they water 1 , ground 2 , air 3 , or space 4 . In the last decades, part of the engineering community has sought inspiration from animals 2,5,6 . For example, flying insects such as bees and flies share the same requirements as light-weight flying vehicles manoeuvring in various habitats from almost object-free terrains 7 to overly cluttered forests 8 via human-made landscapes.…”

Section: Introductionmentioning

confidence: 99%

“…To date, most of the mimics of the collision avoidance machinery rely on traditional cameras from which every pixel at every time point (i.e. at a fixed sampling frequency) needs to be processed 6,15,16,[22][23][24] . The processing occurs even when nothing is changing in the agent's surroundings.…”

Section: Introductionmentioning

confidence: 99%

Finding the Gap: Neuromorphic Motion Vision in Cluttered Environments

Schoepe

Janotte

Milde

et al. 2022

Preprint

View full text Add to dashboard Cite

Many animals meander in environments and avoid collisions. How the underlying neuronal machinery can yield robust behaviour in a variety of environments remains unclear. In the fly brain, motion-sensitive neurons indicate the presence of nearby objects and directional cues are integrated within an area known as the central complex. Such neuronal machinery, in contrast with the traditional stream-based approach to signal processing, uses an event-based approach, with events occurring when changes are sensed by the animal. Contrary to classical von Neumann computing architectures, event-based neuromorphic hardware is designed to process information asynchronously and in a distributed manner. Inspired by the fly brain, we model, for the first time, a neuromorphic closed-loop system mimicking essential behaviours observed in flying insects, such as meandering in clutter and crossing of gaps, both of which are also highly relevant for autonomous vehicles. We implemented our system both in software and on neuromorphic hardware. While moving through an environment, our agent perceives changes in its surroundings and uses this information for collision avoidance. The agent's manoeuvres result from a closed action-perception loop implementing probabilistic decision-making processes. This loop-closure is thought to have driven the development of neural circuitry in biological agents since the Cambrian explosion. In the fundamental quest to understand neural computation in artificial agents, we come closer to understanding and modelling biological intelligence by closing the loop also in neuromorphic systems. As a closed-loop system, our system deepens our understanding of processing in neural networks and their computations in both biological and artificial systems. With these investigations, we aim to set the foundations for neuromorphic intelligence in the future, moving towards leveraging the full potential of neuromorphic systems.

show abstract

Insect-inspired vision for autonomous vehicles

Cited by 14 publications

References 57 publications

Mechanisms of punctuated vision in fly flight

Mechanisms of punctuated vision in fly flight

Comparison of Deep Learning Techniques for Classification of the Insects in Order Level With Mobile Software Application

Finding the Gap: Neuromorphic Motion Vision in Cluttered Environments

Contact Info

Product

Resources

About