Transcript Diapositiva 1 - a4academics.com

Seminar on,
Applications of Artificial
Intelligence in Safe
Human-Robot Interactions
Contents
 Introduction
 Human Modelling
 Prediction of the Human Trajectory
 Reactive Control Scheme
 Conclusion
 References
2
Introduction
Integration of both Robot’s and Human’s
workspace.
A new sensory system for
modelling,tracking & predicting human
motions within a Robot workspace
Obtain a superquadric-based model of
human using SOM.
Assess the danger of the robot operations.
A new reactive control scheme.
3
Human modelling
Sensory system for modelling and tracking human
motion.
Safety mat.
Data processing using SOM.
Obtain body orientation and location.
Model of the human body.
4
Four steps in Human modelling
Step 1: The safety mat consists of a number of pressureactivated nodes.Each node on the mat has fixed
coordinates. Under the human body weight, a set of
nodes F = {(xj, yj)|j = 1, . . . , n} are activated across the
mat.
Step 2:This set is then clustered into two subsets F1 and
F2, corresponding to each foot using a SOM network .
Step 3: Using these subsets, the human body orientation
and its location are derived .
Step 4: This information along with average human body
dimensions is then used in order to obtain a model of the5
human .
Safety mat
Detects obstacles.
Constructed using 2 rubber sheets having
parallel wires.
Pressure activated nodes.
Each node on the mat has fixed
coordinates.….
6
Data processing using SOM
Activated node set F needs to be first divided(clustered) into two
subsets,F1 and F2 corresponding to each foot.
SOM network seems a suitable candidate for clustering the data
representing human footprints.
Input to the SOM network is (xj, yj) pairs and produce output as
(f1, f2)
7
Data processing using SOM
Correct clustering
TYPE A
Incorrect clustering
TYPE B
8
Connectivity of the Laplacian matrics of the
given sample set.
Type A sample set have 2 zero eigen values in its Laplacian matrics…
9
 To convert type B to type A, uncertain nodes are need to
be removed.
 l1 and l2 corresponding to the outer borders and
orientation of each soles.
 lavg represents inner border of the two soles.
10
Obtaining body orientation
It can be obtained from 2 subsets F1 & F2
 α- average of sole orientation
Centre of the body
These values used to obtain human model
Lines lL and lR lines connecting the centers of
the forefoot to the heel in each sole, respectively.
11
Model of the human body
Unduloidlike structure
Variable cross section at various heights.
12
13
Prediction of the human
trajectory
Motivated by ordinary human-human
interaction.
Observe the pattern of the motion and
predict the motion using ANN.
14
15
Reactive control
Formulation of the Danger Index.
DI(κ, v) = fD(κ)fV(v)
Impedence based Reactive Control
Strategy.
-Threshold value for DI.
-Repulsive force
-Virtual damping torque.
Cumulative PDI method
Fp = kpPDI(n0, dp)up
16
17
Conclusion
Study on a sensory system and reactive
control scheme.
SOM network and superquadric functions
are used for human modelling.
Human motion predicted using ANN.
Reactive control scheme is developed.
18
References
 Nima Najmaei and Mehrdad R. Kermani “Applications of
Artificial Intelligence in Safe Human–Robot Interactions”,
April 2011.
 A. De Santis, B. Siciliano, A. De Luca, and A. Bicchi, “An
atlas of humrobot interaction,” Mech. Mach. Theory, vol.
43, pp. 253–270, 2008.
 R. Bischoff and V. Graefe, “Hermes—A versatile
personal robotic assistant,” Proc. IEEE, vol. 92, no. 11,
pp. 1759–1779, Nov. 2004.
 American Nat. Standard for Indus. Robots—Safety
Requirement, RIA/ANSI R15.06—1999, 1999.
19