Abstract
We demonstrate a hybrid neuromorphic learning paradigm that learns complex sensorimotor mappings based on a small set of hard-coded reflex behaviors. A mobile robot is first controlled by a basic set of reflexive hand-designed behaviors. All sensor data is provided via a spike-based silicon retina camera (eDVS), and all control is implemented via spiking neurons simulated on neuromorphic hardware (SpiNNaker). Given this control system, the robot is capable of simple obstacle avoidance and random exploration. To train the robot to perform more complex tasks, we observe the robot and find instances where the robot accidentally performs the desired action. Data recorded from the robot during these times is then used to update the neural control system, increasing the likelihood of the robot performing that task in the future, given a similar sensor state. As an example application of this general-purpose method of training, we demonstrate the robot learning to respond to novel sensory stimuli (a mirror) by turning right if it is present at an intersection, and otherwise turning left. In general, this system can learn arbitrary relations between sensory input and motor behavior.
Highlights
A long-standing dream for robotics is to provide the same sort of intelligent, adaptive, and flexible behavior that is seen in living biological systems
The algorithm described in this paper is a general-purpose approach to implementing mobile robot behaviors, making use of massively parallel low-power neuromorphic computing hardware
The basic algorithm is as follows: 1. define a set of basic actions; 2. define functions for the strength of each action, given the current sensory state; 3. generate a neural network that approximates these functions; 4. use the neural network to drive the robot and record the neural activity; 5. identify a set of positive examples where the robot has accidentally performed whatever other task is desired; 6. retrain the neural connections between the sensory neurons and the action strengths so as to approximate the data gathered during the positive examples
Summary
A long-standing dream for robotics is to provide the same sort of intelligent, adaptive, and flexible behavior that is seen in living biological systems. By creating systems that emulate biological adaptivity, we can investigate intelligence in a very broad sense, including capabilities that are not yet seen in current machine learning methods (McFarland and Bösser, 1993). It is still an open research question as to how to use biological inspiration to construct improved methods of robotic control. The hardware used is standard generalpurpose computing, and the algorithms are machine-learning based This means that, while they may take some high-level inspiration from biology, the algorithms themselves do not directly map to biological details
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
