Return to Job offers

Post-doc project : Application-aware wake-up/stand-by control in the reconfigurable platform L-iot

— Environment (Inria team, location, contacts):

Eric Rutten
Ctrl-A team, INRIA Grenoble – Rhone-Alpes,
Antenne GIANT, Batiment 50C Minatec, 17 rue des Martyrs 38054 Grenoble Cedex

in cooperation with

— Topic:

Application-aware wake-up/stand-by control in the reconfigurable platform L-iot

In the Internet of Things (IoT), devices have more and more communication and computation capabilities, while requiring extended autonomy, hence efficient energy management.
CEA LETI is developing the reconfigurable platform L-iot, equipped with mechanisms for putting different subsystems (processing, radio communication, specialized co-processors, sensors) to sleep mode, and switching them on again from stand-by modes by a wake-up controller.
Switching between these partially on and off configurations is controlled in response to events and measures from internal states of the architecture and battery, or external events from sensors, but it also has to take into account the ongoing computations and functionalities, not always interruptible, as well as higher-level application objectives.
The control has to fulfill complex objectives integrating functionality fulfilling, energy sparing, and degraded modes management.
A typical use-case considered at CEA DACLE/LIALP is that of a smart wearable/portable spatial exploration system (called hereafter INSPEX system) that can typically be integrated in a white cane, as envisioned in the H2020 ICT-3 project INSPEX (nb 730953). The INSPEX main system consists of several range sensor technologies, some of them being quite power hungry, associated with an Inertial Measurement Unit in order to build a map of the user’s surrounding and detect obstacles. The system is immersed in smart environments, being part of the IoT. In order to better deal with its power consumption, it might highly benefit from the L-IoT architecture, especially by an stringent power management strategy that will switch on/off parts of the system, depending on external event and/or on available energy

The topic of this proposal is to work on design methods for such application-aware wake-up/stand-by controller, following approches based on behavioral models and control techniques. The goal is to obtain controllers that ensure safety of the system executions, and correct fulfilment of the objectives.
In the general context of Autonomic Computing, explicitly designed feedback loops are performing the management of adaptations at hardware and software levels. To go beyond the difficult and error-prone hand writing of reaction rules, with risks of inconsistencies or incomplete management, we will adopt controller design methods based on models of the dynamics of the platform.

We will explore and propose design methods for the analysis and modeling of the configurations space of the platform and the applications, the specification of applicative as well as platform control objectives, and the design of controllers enforcing them. We will consider automata based models for the space of configurations (e.g., avoidance of state from which the system can not be woken up), and use tools based on reactive languages developed in the Inria Ctrl-A team. We will also consider continuous models for quantitative aspects of the system (e.g. energy related) or stochastic models for the management of uncertainty. We will also consider the coordination of these different loops. The proposed methods will be validated experimentally on use cases from the automated cane application.

— Skills:

experience in some of the following : reconfigurable architectures, infrastructrues of the IoT, autonomic computing, reactive languages.

References :

UTBB FDSOI technology flexibility for ultra low power internet-of-things applications
E. Beigne, J. F. Christmann, A. Valentian, O. Billoint, E. Amat and D. Morche
45th European Solid State Device Research Conference (ESSDERC), Graz, 2015, pp. 164-167.

Model-based design of correct controllers for dynamically reconfigurable architectures
Xin An, Eric Rutten, Jean-Philippe Diguet, Abdoulaye Gamatié
ACM Transactions on Embedded Computing Systems (TECS), ACM, 2016, 15 (3), pp.#51. <10.1145/2873056>

— Key Words:

reconfigurable architectures, infrastructrues of the IoT, autonomic computing, reactive languages, system reconfiguration