A neural network and IoT-based scheme for performance assessment in Internet of Robotic Things

Abstract : Internet of Robotic Things (IoRT) is a new concept introduced for the first time by ABI Research. Unlike the Internet of Things (IoT), IoRT provides an active sensorization and is considered as the new evolution of IoT. This new concept will bring new opportunities and challenges, while providing new business ideas for IoT and robotics’ entrepreneurs. In this paper, we will focus particularly on two issues: (i) connectivity maintenance among multiple IoRT robots, and (ii) their collective coverage. We will propose (i) IoRT-based, and (ii) a neural network control scheme to efficiently maintain the global connectivity among multiple mobile robots to a desired quality-of-service (QoS) level. The proposed approaches will try to find a trade-off between collective coverage and communication quality. The IoT-based approach is based on the computation of the algebraic connectivity and the use of virtual force algorithm. The neural network controller, in turn, is completely distributed and mimics perfectly the IoT-based approach. Results show that our approaches are efficient, in terms of convergence, connectivity, and energy consumption.
Complete list of metadatas

Cited literature [22 references]  Display  Hide  Download

https://hal.inria.fr/hal-01261842
Contributor : Jean Cristanel Razafimandimby <>
Submitted on : Wednesday, April 6, 2016 - 9:35:29 AM
Last modification on : Thursday, April 7, 2016 - 1:06:17 AM
Long-term archiving on : Thursday, July 7, 2016 - 4:41:53 PM

File

Cristanel-Razafimandimby-paper...
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01261842, version 1

Collections

Citation

Cristanel Razafimandimby, Valeria Loscri, Anna Maria Vegni. A neural network and IoT-based scheme for performance assessment in Internet of Robotic Things . I4T - 1st International Workshop on Interoperability, Integration, and Interconnection of Internet of Things Systems, Apr 2016, Berlin, Germany. ⟨hal-01261842⟩

Share

Metrics

Record views

376

Files downloads

3853