The emergence of collective obstacle avoidance based on a visual perception mechanism

Qi, Jingtao; Bai, Liang; Xiao, Yandong; Wei, Yingmei; Wu, Wansen

doi:10.1016/j.ins.2021.10.039

Cited by 17 publications

(16 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We took the classic and well-studied approach of modelling generic agent behaviour with abstract zonal models, since our main goal was to get an understanding of general mechanisms enabling escape from difficult obstacles rather than to model specific animals or scenarios. However, a variety of mechanistically different models which do not use radial relations [13,15], or even agent-to-agent interactions at all [15][16][17][18]69], can also lead to effective swarming behaviour which may fit empirical data as well or better than the radius models. Even for radius-based models, there is some discussion in the literature of the nature and necessity of alignment in swarms, since the explicit alignment of orientation with agents can be replaced with matching the velocity vectors of other individuals [73].…”

Section: Discussionmentioning

confidence: 99%

“…Observational studies tracking members of animal groups in their natural habitats typically only investigate the large-scale impacts of environmental features on dispersal and habitat use, with limited insights into the immediate response of animal groups to obstacles [9,34,35,41]. The few laboratory and theoretical works on collective movement around obstacles usually only consider obstructions like arena boundaries or small, convex objects [10,11,18,39,[42][43][44][45]. Interactions between animal swarms and predators have been explored more [8,30,[46][47][48], but while behaviours for predator avoidance may overlap with navigation around obstacles, there are clear, important differences in these contexts.…”

Section: Introductionmentioning

confidence: 99%

“…Several decades of theoretical and empirical research have resulted in valuable insights into collective motion, including a clear understanding of types of individual behavioural rules that produce global dynamics similar to what we observe in natural swarms [13–18]. Among the most widely used types of models [7] are those based on behavioural rules determined by the distances to and angles of nearby neighbours.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Alignment with neighbours enables escape from dead ends in flocking models

Joshi

Popp

Werfel

et al. 2022

J. R. Soc. Interface.

View full text Add to dashboard Cite

Coordinated movement in animal groups (flocks, schools, herds, etc.) is a classic and well-studied form of collective behaviour. Most theoretical studies consider agents in unobstructed spaces; however, many animals move in often complicated environments and must navigate around and through obstacles. Here we consider simulated agents behaving according to typical flocking rules, with the addition of repulsion from obstacles, and study their collective behaviour in environments with concave obstacles (dead ends). We find that groups of such agents heading for a goal can spontaneously escape dead ends without wall-following or other specialized behaviours, in what we term ‘flocking escapes’. The mechanism arises when agents align with one another while heading away from the goal, forming a self-stable cluster that persists long enough to exit the obstacle and avoids becoming trapped again when turning back towards the goal. Solitary agents under the same conditions are never observed to escape. We show that alignment with neighbours reduces the effective turning speed of the group while letting individuals maintain high manoeuvrability when needed. The relative robustness of flocking escapes in our studies suggests that this emergent behaviour may be relevant for a variety of animal species.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Alignment with neighbours enables escape from dead ends in flocking models

Joshi

Popp

Werfel

et al. 2022

J. R. Soc. Interface.

View full text Add to dashboard Cite

show abstract

“…The vision-based model of collective motion presented by Bastien et al [29] and extended to obstacle avoidance by Qi et al [30] offers a different approach for perception modeling. They introduce specifically designed complex integro-differential operators that, depending on the visual input, govern an agent's motion, which in turn creates various collective behaviors.…”

Section: Introductionmentioning

confidence: 99%

“… / 10 [BL], respectively (agent BL[20][21][22][23][24][25][26][27][28][29][30]. Data points are the mean order parameter of the swarm at each simulation frame, with standard error margin.…”

mentioning

confidence: 99%

Vision-Based Collective Motion: A Locust-Inspired Reductionist Model

Krongauz

Ayali²,

Kaminka³

2023

Preprint

View full text Add to dashboard Cite

Collective motion is an omnipresent, fascinating phenomenon. Swarming individuals aggregate, coordinate and move, utilizing only local social cues projected by conspecifics in their vicinity. Major theoretical studies assumed perfect information availability, where agents rely on complete and exact knowledge of inter-agent distances and velocities. However, the sensory modalities that are responsible for the acquisition of environmental information were often ignored. Vision plays a central role in animal perception, and in many cases of collective motion, it is the sole source of social information. Here we investigate a vision-based collective motion model inspired by locust marching bands. We address two major challenges: the evaluation of distance and velocity and visual occlusions. We compare three strategies an agent can use to interpret partially occluded visual information. These differ in the visual cognition capabilities of the agent and the respective computational requirements.In silicoexperiments conducted in various geometrical conditions show that the three approaches display different rates of convergence to an ordered state: The least computationally-demanding approach, in which no peer recognition takes place, shows slower convergence in geometrically-constrained environments. Our findings provide insights into the visual processing requirements from biological as well as artificial swarming agents in complex settings.

show abstract

Hardware Implementation of Ultra‐Fast Obstacle Avoidance Based on a Single Photonic Spiking Neuron

Gao,

Xiang,

Song

et al. 2023

Laser & Photonics Reviews

View full text Add to dashboard Cite

Visual obstacle avoidance is widely applied to unmanned aerial vehicles (UAVs) and mobile robot fields. A simple system architecture, low power consumption, optimized processing, and real‐time performance are extremely needed due to the limited payload of some mini UAVs. To address these issues, an obstacle avoidance system harnessing the rate encoding features of a photonic spiking neuron based on a Fabry–Pérot (FP) laser is proposed, which simulates the monocular vision. Here, time to collision is used to describe the distance of obstacles. The experimental results show that the FP laser excites ultra‐fast spike responses in real time for the following cases, facilitating the generation of control commands by motor neurons to realize accurate decision‐making. Four cases of mobile obstacle avoidance scenarios, including “Constant velocity approach”, “Approach and retreat”, “The motion state involving stays”, and “Approach with different velocities”, and obstacle avoidance problems with multiple stationary obstacles appearing simultaneously are experimentally analyzed. The system exhibits a spike response rate of up to 5 GHz. This work proves the feasibility of applying the ultra‐fast photonic obstacle avoidance system to UAVs and other fields in the future and highlights the potential of photonic neuromorphic processor platforms.

show abstract

The emergence of collective obstacle avoidance based on a visual perception mechanism

Cited by 17 publications

References 41 publications

Alignment with neighbours enables escape from dead ends in flocking models

Alignment with neighbours enables escape from dead ends in flocking models

Vision-Based Collective Motion: A Locust-Inspired Reductionist Model

Hardware Implementation of Ultra‐Fast Obstacle Avoidance Based on a Single Photonic Spiking Neuron

Contact Info

Product

Resources

About