Autonomous Bio-Inspired Small-Object Detection and Avoidance

Ohradzansky, Michael; Alvarez, Hector E; Keshavan, Jishnu; Ranganathan, Badri N.; Humbert, J. Sean

doi:10.1109/icra.2018.8461156

Cited by 10 publications

(11 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Insects rely on patterns of optic flow to generate a centering flight response (Srinivasan et al, 1998), which was tested in (Santos-Victor and Sandini, 1997) and (Griffiths et al, 2006), whereas our systems rely on LiDAR depth measurements. Insects also rely on optic flow for small-obstacle detection and avoidance (Alvarez et al, 2019), and which has been demonstrated as a viable obstacle detection method onboard multi-rotor platforms (Ohradzansky et al, 2018). For the Tunnel circuit event, our team deployed a fleet of ground vehicles in a largely 2D environment.…”

Section: Metric-topological Planning and Reactive Controlmentioning

confidence: 99%

“…Tuning constraints on such a problem is often non-trivial and highly dependent on the environment. The concept of planning, based on direct high resolution depth information, has been explored in past in the context of bio-inspired obstacle avoidance (Ohradzansky et al, 2018;Alvarez et al, 2019;Ohradzansky et al, 2020) and lookahead planning within a depth image (Matthies et al, 2014;Dubey et al, 2017;Ahmad and Fierro, 2019). These methods, however, are either only suitable for 2D navigation or are based on lookahead planning.…”

Section: Aerial Vehicle Vision-based Local Controlmentioning

confidence: 99%

See 1 more Smart Citation

Multi-Agent Autonomy: Advancements and Challenges in Subterranean Exploration

Ohradzansky¹,

R.²,

Riley³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Artificial intelligence has undergone immense growth and maturation in recent years, though autonomous systems have traditionally struggled when fielded in diverse and previously unknown environments. DARPA is seeking to change that with the Subterranean Challenge, by providing roboticists the opportunity to support civilian and military first responders in complex and high-risk underground scenarios. The subterranean domain presents a handful of challenges, such as limited communication, diverse topology and terrain, and degraded sensing. Team MARBLE proposes a solution for autonomous exploration of unknown subterranean environments in which coordinated agents search for artifacts of interest. The team presents two navigation algorithms in the form of a metric-topological graph-based planner and a continuous frontier-based planner. To facilitate multi-agent coordination, agents share and merge new map information and candidate goal-points. Agents deploy communication beacons at different points in the environment, extending the range at which maps and other information can be shared. Onboard autonomy reduces the load on human supervisors, allowing agents to detect and localize artifacts and explore autonomously outside established communication networks. Given the scale, complexity, and tempo of

show abstract

Section: Metric-topological Planning and Reactive Controlmentioning

confidence: 99%

Section: Aerial Vehicle Vision-based Local Controlmentioning

confidence: 99%

Multi-Agent Autonomy: Advancements and Challenges in Subterranean Exploration

Ohradzansky¹,

R.²,

Riley³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Through observation of insects and other flying animals in different environments, optic flow-based navigation methods were developed [ 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 ]. Small-obstacle avoidance algorithms based on optic flow measurements are demonstrated onboard a multirotor platform in [ 43 , 44 ].…”

Section: Introductionmentioning

confidence: 99%

Lidar-Based Navigation of Subterranean Environments Using Bio-Inspired Wide-Field Integration of Nearness

Ohradzansky

Humbert

2022

Sensors

Self Cite

View full text Add to dashboard Cite

Navigating unknown environments is an ongoing challenge in robotics. Processing large amounts of sensor data to maintain localization, maps of the environment, and sensible paths can result in high compute loads and lower maximum vehicle speeds. This paper presents a bio-inspired algorithm for efficiently processing depth measurements to achieve fast navigation of unknown subterranean environments. Animals developed efficient sensorimotor convergence approaches, allowing for rapid processing of large numbers of spatially distributed measurements into signals relevant for different behavioral responses necessary to their survival. Using a spatial inner-product to model this sensorimotor convergence principle, environmentally relative states critical to navigation are extracted from spatially distributed depth measurements using derived weighting functions. These states are then applied as feedback to control a simulated quadrotor platform, enabling autonomous navigation in subterranean environments. The resulting outer-loop velocity controller is demonstrated in both a generalized subterranean environment, represented by an infinite cylinder, and nongeneralized environments like tunnels and caves.

show abstract

“…Examples include depth maps (distance measurements per pixel), optic flow data (motion patterns between consecutive frames, in pixel space) or semantic maps (which describe for each pixel the object it belongs to). Inputs such as these are often used in bio-inspired computer vision applications for navigation [18,14,22,1], and can provide novel insight into bird behaviour. We demonstrate the method by apply- ing it to a small sample of flights of a Harris' hawk executing perching and obstacle avoidance manoeuvres, which is part of a larger dataset that will be analysed fully elsewhere.…”

Section: Introductionmentioning

confidence: 99%

Through hawks’ eyes: reconstructing a bird’s visual field in flight to study gaze strategy and attention during perching and obstacle avoidance

Miñano

2021

Preprint

View full text Add to dashboard Cite

We present a method to analyse visual attention of a bird in flight, that combines motion capture data with renderings from virtual cameras. We applied it to a small subset of a larger dataset of perching and obstacle avoidance manoeuvres, and studied visual field stabilisation and gaze shifts. Our approach allows us to synthesise visual cues available to the bird during flight, such as depth information and optic flow, which can lead to novel insights into the bird's gaze strategy in flight. This preliminary work demonstrates the method and suggests several new hypotheses to investigate with the full dataset.

show abstract

Autonomous Bio-Inspired Small-Object Detection and Avoidance

Cited by 10 publications

References 25 publications

Multi-Agent Autonomy: Advancements and Challenges in Subterranean Exploration

Multi-Agent Autonomy: Advancements and Challenges in Subterranean Exploration

Lidar-Based Navigation of Subterranean Environments Using Bio-Inspired Wide-Field Integration of Nearness

Through hawks’ eyes: reconstructing a bird’s visual field in flight to study gaze strategy and attention during perching and obstacle avoidance

Contact Info

Product

Resources

About