different robots to test the validity of computing cluster in our
robotic service cloud. The size of the data is carried 500MB
and 1GB respectively with one, two and three computing
nodes cluster that uses map-reduce algorithm to process the
data as shown in Fig 5 and 6. The performance graph clearly
indicates the usefulness of map-reduce computing cluster
which makes computation faster depending upon the size of
the cluster. In the graphs, time(in seconds) taken by different
size of cluster is shown and it depends upon the computational
power of individual unit present in the cluster. Hence, the performance
will vary when experimenting on different hardware
configuration.
IV. CONCLUSION AND FUTURE WORK
In this paper, we have shown an approach to introduce
robotic services in cloud. At this cloud a client can have
services like navigation, map building, object recognition etc.
Moreover, a rapidly executed service can be obtained with
the use of map-reduce computing cluster. The services are
provided with the use of visual programming language of
MRDS. And some results of map-reduce computing cluster
are also presented. The whole system is tested successfully
for speech based navigation using Create robot. In this work,
we have enlisted only few services whereas, more number
of services can be added to improve functionality of the
overall system. Furthermore, we have to include support for
other robots with different structured/unstructured working
environment. But, this work opens the scope for virtualization
of robots by offering different robotic services and a test-bed
for testing robotic algorithms.