Startup creation: W4I: Wireless For Industries
W4I Offering a new generation of Industrial Internet of Things: industrial sensors are wired to nodes of a high-bitrate wireless network, providing autonomous functionality with wide spatial coverage (10000m2/node). Enhanced with dedicated software, these solutions are aimed at addressing monitoring and mitigating the impact of crises (critical air pollution), digitizing operational theaters (floods, fires, gas), and managing both the air and energy of buildings.
Context
Since January 2021, Members of ROCS have been working with the Feder Wizard project team to mature the research outcomes at the end of the project. These outcomes involve deploying industrial sensors to collect data and transmit them through a mesh wireless network. This maturation process will lead to the creation of a new startup to commercialize products for industrial site markets (gas, electricity, water...), tourist sites (sea pollution, forest fires, rising water levels in canyons...), and hazardous locations (Seveso sites, transportation of hazardous materials...).
The creation project has been selected by the CNRS Innovation RISE program for support until the company is established.
Contribution
The elements contributed to the project of start-up creation are:
- Excelent results of our research work in IoT
- Competent team members in the maturation process
- Excellent collaboration between the different entities: University of Corte, University of Paris-Saclay and CNRS
Impact
The technological impacts of the solution are:
- Creation of Edge technology for Industrial IoT.
- Reduction of the IoT Carbon Footprint.
- Providing solutions for data sovereignty.
The societal impacts are:
- Transferring research results to the industry.
- Job creation.
Article
F. Quezada, C. Gicquel, S. Kedad-Sidhoum. Combining polyhedral approaches and stochastic dual dynamic integer programming for solving the uncapacitated lot-sizing problem under uncertainty. INFORMS Journal on Computing, 2022, vol. 34, issue, pp 1024–1041. DOI 10.1287/ijoc.2021.1118 - HAL hal-03606367
Context
This paper investigates a combinatorial optimization problem arising in the context of the operational management of an industrial production site. More specifically, this problem deals with deciding which to produce in the production site, when and each quantity, for a planning horizon typically covering a few days or weeks. The paper considers a stochastic extension of this problem in which the input parameters are subject to uncertainty and models it as a multistage stochastic integer program.
Multi-stage stochastic integer programs are very challenging to solve and there is currently a large gap between the size of the actual production planning problems encountered in industry and the size of the problems that can be solved by the current state-of-the-art algorithms. The main objective of this work was to contribute in closing this gap by investigate how a recently published algorithm could be adapted and extended to solve production planning and lot-sizing problems.
Contribution
The main research contribution in this paper consists in proposing a novel extension of the stochastic dual dynamic integer programming (SDDiP) algorithm in order to make it more computationnally efficient at solving lot-sizing problems.
Impact
This paper has been cited 8 times according to Google Scolar. This work was presented in various workshops and conferences, including an invited talk within the virtual seminar series of the Stochastic Programming Society.
Article:
D. Mishra, N. R. Zema and E. Natalizio A High-End IoT Devices Framework to Foster Beyond-Connectivity Capabilities in 5G/B5G Architecture. in IEEE Communications Magazine, vol. 59, no. 1, pp. 55-61, January 2021, DOI 10.1109/MCOM.001.2000504 - HAL hal-03541400.
Context
The concept of the Internet of Things (IoT) is evolving rapidly, encompassing several application domains including Internet of Robotic Things (IoRT), Internet of Autonomous Things (IoAT), Internet of Space Things (IoST or CubeSats), Internet of Mobile Things (IoMT), and so on. Such a diverse notion of “things” (i.e., Internet of Everything or IoX) faces new challenges in terms of convergence, discoverability, programmability, and communication.
Although system innovations like software defined networking (SDN), network functions virtualization (NFV), and mobile edge computing (MEC) exist to acquire the desired level of service abstraction over a common underlying physical infrastructure, there is still a clear gap to bridge for seamless integration and deployment of smart devices into 5G because of characterization disparity of the device capabilities and perceived intelligence.
Contribution
Design of a unified and coherent software architecture involving functions, mechanisms, algorithms, and protocols that will assist in seamless integration and management of these high-end devices still remains a challenge. This work aims at filling this gap by providing a software framework consisting of a reference architecture, a network suite, and a set of tailored mechanisms for seamless integration of high-end devices toward realizing a vision of IoX ecosystems.
It is important to present key distinctions of the proposal, IoT High-end Autonomous CooperAtive framework (ITHACA), from the MEC framework that enables readers to differentiate the two paradigms. MEC offers content caching or offloading of a compute-intensive task to cloud in MEC for saving device-bound resources (onboard energy, communication latency, bandwidth, proximity, etc.) instead of exploiting cognitive intelligence and cooperative ability of smart devices. ITHACA is bound with the ability of the architecture to trigger algorithms and services to seamlessly integrate a device and to opportunistically leverage what a device is able to perform.
Impact
This work was published in a top tier IEEE Magazine (IEEE Communication Magazine, IF: 11) and is the fullfilment of a strong international collaboration (between France and U.A.E.) which is leading to common international research projects and PhD co-supervision.
Article
H. De Oliveira , M. Kaneko, L. Boukhatem, E. Hidemi Fukuda Deep Reinforcement Learning-Aided Optimization of Multi-Interface Allocation for Short-Packet Communications. IEEE Transactions on Cognitive Communications and Networking, pp. 738-753, Vol. 9, No. 3, June 2023. DOI 10.1109/TCCN.2023.325266 - HAL hal-03944071v1
Context
The severe spectrum scarcity and the stringent requirements of Beyond 5G applications call for an integrated use of low frequency Sub-6 GHz and high frequency millimeter Wave bands. Focusing on future Internet of Things (IoT) Short-Packet Communications (SPC), this work investigates the optimized usage of such diverse wireless interfaces (namely Sub-6 GHz and mmWave bands). To solve this intricate problem, we propose a two-stage resolution framework that combines the advantages of DRL (Deep Reinforcement Learning) and DCP (Difference of Convex Programming) optimization methods. We develop an unifying framework devoted to SPC that jointly optimizes the user partitioning over each band, and the radio resource scheduling within each band.
Contribution
The proposed framework enables to efficiently tackle the challenges imposed by dynamically varying mobile environments such as the Line-of-Sight situations of the wireless links, and the heterogeneity of individual users Quality of Service (QoS) requirements, such as rate, delay and reliability.
The work has two major novelties : to reduce the time complexity of the resource allocation problem using both DCP and regularization while guaranteeing convergence to a local optimum, and more generally prove the capability and effectiveness of artificial intelligence tools to solve complex real-time radio resource allocation problems.
Impact
This work was published in a top tier IEEE journal and is the fullfilment of a strong international collaboration (between France and Japan) which has led to common international research projects and PhD co-supervision.