The Centre for Artificial Intelligence and Robotics (CAIR) was
established on 17 October 1986, as an R&D lab under the Defence
Research and Development Organisation (DRDO). Situated at
Bangalore, the charter of duties of CAIR is to conduct basic and applied
research in the fields of information sciences, artificial intelligence,
robotics and automation with emphasis on defence requirements. The thrust
areas of CAIR are as follows:
Artificial Intelligence
Robotics
Neural Networks
Computer Vision
Virtual Reality
Control Systems
Signal Processing
Embedded Systems
Communication and Information
Security
Tactical Communication Systems
Command & Control Systems
CAIR is
unique in combining a wide variety of activities from the theoretical to the
applied in a broad spectrum of areas. The aim of CAIR is to create an
intellectually stimulating atmosphere with excellent computing and library
facilities. CAIR is a 200-strong lab with equal emphasis on technology
development and state-of-art systems development.
Artificial Intelligence
Pilot Associate Kernel: It is a toolbox for developing concurrent
hierarchical applications with real-time constraints. It is developed on a Sun
Spark station using posix with real-time features of the Solaris operating
system. It provides features for event-based programming in a multitasking
environment. It also provides methods for control, communication and
sychronisation between tasks.
Intelligent Warning Manager: For managing and displaying warnings to the pilot.
The warning manager reasons with the set of warnings and displays only those are
important to the pilot based on the warning knowledge provided to the warning
manager.
Robot Path Planning: Software for finding a 2-dimensional path for a
mobile robot with static obstacles, which can have arbitrary (nonconvex)
polygonal shapes. The path planning is done in the configuration space using a
visibility graph and through free space decomposition techniques.
Network Security System: For service authentication a wrapper program has
been used to facilitate transparent service (e.g. telnet, ftp etc) and host (IP
number, host name) based authentication before providing/denying the service.
For clear text transmission over the network, For countering threats,
stream-based modules have been implemented to encrypt complete network sessions
such as telnet, rlogin etc.
Controllers and Drovers for Robots: A PC based controller using a motion controller IC
(HCTL 1100) has been developed. Another controller in which the closed loop
control is at the PC level has been developed. Besides this, electric drives for
DC and stepper motors providing the appropriate power levels to the motors to be
used for robotic application has been developed.
Intelligent Wheel Chair: Developed for physically handicapped persons. Three
different configurations, viz.
A voice
operated wheelchair
Intelligent
wheelchair which utilizes a model of the room (built into its knowledge base for
path planning.
A camera
system to get information about the room
Information Theory: Research has continued into the encoding and
propagation of information in neural networks leading to the formulation of a
new algorithm for designing neural networks.
Hardware for Neural Networks: The emphasis of neural networks hardware
development at CAIR is on demonstrating the feasibility of using off-the-shelf
DSP hardware for building neural networks. A single layer of a neural network
involving 1024 neurons with 256 inputs each has been designed and tested.
Statistical Physics Methods: Neural networks models have been given a systematic
and theoretical basis in the space of interactions approach of Gardner et al,
using statistical mechanics techniques
In Flight Simulation-2 Programme of the LCA: The USAF F-16 Variable Stability Research Aircraft
(VISTA) operated by Calspan Corporation. Buffalo, USA was used to simulate the
LCA control laws in flight. This in-flight simulation session was primarily used
to evaluate the LCA control laws for the up and away flight phase. The tasks
used to evaluate the control laws included close formation flying, air-to-air
tracking the general manouvers.
Design of Structural Notch Filters Based on SCT-1
Experiments: Designed
notch filters based on the structural coupling tests (SCT-1) on LCA
Autopilot Control Laws for LCA (TD-2): Study on the Autopilot control laws for L|CA Technology Demonstrator 2 (TD-2) phase was carried out.