1Beam Auto Pilot.
Anil Kumar Bheemaiah, Synergy Robotics, AB [email protected],
8/24/2018 1.31 PM
Abstract
This paper describes the algorithms and models for a BEAM auto-pilot design with a monocular camera and microphone as sensors, it also describes the creation of an earth mirror using infrasonics for the mapping of stationary and moving objects. Optionally a SLAM and an auto navigation system can be designed using the Earth Mirror(™) and is the topic of a future publication.
Key Words: BEAM , MFA I, Earth Mirror, OS , autonomous systems, autopilot, analog circuits, behaviour driven designs,
1 Earth Mirror( ) is an integration, non available in combat or strategic use cases for autopilot systems, ™ necessitating only civilian use cases. It is hence not a dual use technology.
Introduction design choices the builder makes while implementing the desired functionality. BEAM Robotics
BEAM robots(Contributors to Wikimedia projects 2003)(Wikipedia 2013; Hrynkiw Unlike many other types of robots and Tilden 2002) may use a set of the controlled by microcontrollers, BEAM analog circuits, mimicking biological robots are built on the principle of using neurons, to facilitate the robot's response multiple simple behaviours linked to its working environment. directly to sensor systems with little signal conditioning.
There are a large number of BEAM robots designed to use solar power from small Earthworms: Move using a longitudinal solar arrays to power a "Solar Engine" wave motion... Crawlers: Robots that which creates autonomous robots capable move using tracks or by rolling the of operating under a wide range of robot's body with some sort of lighting conditions. appendage.
The "Solar Engine" circuit, many H-bridge Having no long-term memory, BEAM circuits for small motor control, tactile robots generally do not learn from past sensor designs, and meso-scale robot behaviour. construction techniques have been documented and shared by the BEAM community. One of the most advanced BEAM robots in this vein is Bruce Robinson's Hider, which
has an impressive degree of capability for Being focused on "Reaction-based" a microprocessor-less design. behaviors, BEAM robotics attempts to copy the characteristics and behaviours of biological organisms, with the ultimate U.S. Patent 5,325,031(Tilden 1994) - goal of domesticating these "Wild" robots. Adaptive robotic nervous systems and control circuits there for - Tilden's patent;
A self-stabilizing control circuit utilizing The aesthetics of BEAM robots derive pulse delay circuits for controlling the from the principle "Form follows limbs of a limbed robot, and a robot function" modulated by the particular
incorporating such a circuit; artificial We define a set of images from a camera "Neurons". Z, as sum(I), at 30 f/s
We map these images on to D through S as a pulse train, (PCK or Pulse Code Beam robots are biomimetic and use Key)(Bheemaiah 2019). emergent A.I. sS is the semantic map on D of the sum(I), They are simple analogue circuits, we design behaviors of obstacle inspired by insect nervous systems. We avoidance on sS. consider non biomimetic beam devices for auto navigation. Circuit for obstacle avoidance behavior.
We consider the design of a monocular (Zourntos and Mathai 2007) beam robot auto-pilot, using pcm to encode the camera images and next describe an earth mirror design using a microphone for infrasonic sound.
Beam monocular robotics.
All beam works by an internal mapping of the outer world, for the creation of obstacle avoidance and path or track planning. It is possible to integrate google maps to the beam controls for goal directed behavior. We define a semantic map with a hierarchy of two orders of semantics, primary and secondary in semantic transfer.
ST -> [ ST_Primary, ST_Secondary]
If D is the lattice, on which we define a semantic map S, then pS is primary and sS is secondary. sS maps obstacle avoidance, and pS goal directed behavior through Google map integration.
Ref:(Zourntos and Mathai 2007)
Circuit for integration of sS and pS. sS and pS are unified as a Stability of two or more attractors or in the same attractors in semantic transfer.(Zhu and Grefenstette 2017; Ha, n.d.)
Earth Mirror Auto Pilot.
Consider IR(t), the stream from a microphone with infrasonic bandpass.
We define a lattice L with a primary and secondary semantic map, sM and pM, while the pM is for Google map integration, sM defines segmentation of stationary and moving objects defined by a mapping to OOP persistence.
Conclusions and Future Work. Composition.” Proceedings of ACL 2017, Tutorial Abstracts. We have thus presented a theoretical https://doi.org/10.18653/v1/p17-5003. model and algorithms for a monocular Zourntos, Takis, and N. J. Mathai. 2007. “A camera based BEAM system, which shows BEAM-Inspired Lyapunov-Based Strategy for Obstacle Avoidance and emergent behavior of obstacle avoidance, Target-Seeking.” we have also integrated goal directed http://ieeexplore.ieee.org/stamp/stamp.js p?tp=&arnumber=4282969. behavior using Google Maps, we have supplemented this with an earth mirror for mapping stationary and moving objects analogous to SLAM systems.
Future work is to create auto navigation using the infrasonic Earth Mirror(™).
References
Contributors to Wikimedia projects. 2003. “BEAM Robotics - Wikipedia.” Wikimedia Foundation, Inc. April 25, 2003. https://en.wikipedia.org/wiki/BEAM_robot ics. Ha, Hsin-Yu. n.d. “Integrating Deep Learning with Correlation-Based Multimedia Semantic Concept Detection.” https://doi.org/10.25148/etd.fidc000162. Hrynkiw, David, and Mark W. Tilden. 2002. JunkBots, Bugbots, and Bots on Wheels: Building Simple Robots With BEAM Technology. McGraw Hill Professional. Tilden, Mark W. 1994. Adaptive robotic nervous systems and control circuits therefor. USPTO 5325031. US Patent, filed June 15, 1992, and issued June 28, 1994. https://patentimages.storage.googleapis.c om/2f/b0/45/bf157e282ed78a/US5325031 .pdf. Wikipedia, Source. 2013. Biomorphic Robots: Beam Robotics, Humanoid Robots, Data, Android, Aibo, Pino, Gynoid, Rhex, Asimo, Robonaut, Actroid, Xianxingzhe, Geoff Peterson, Bat. University-Press.org. Zhu, Xiaodan, and Edward Grefenstette. 2017. “Deep Learning for Semantic