scispace - formally typeset
Search or ask a question
Proceedings ArticleDOI

Performance Evaluation of Real-Time Scheduling Heuristics for Energy Harvesting Systems

18 Dec 2010-pp 398-403
TL;DR: This paper compares several scheduling heuristics with the optimal algorithm known as LSA (Lazy Scheduling Algorithm) and reports results of an experiment study in terms of percentage of deadlines satisfied.
Abstract: Energy constrained systems can increase their usable lifetimes by extracting energy from their environment. This is known as energy harvesting. This paper investigates scheduling issues in uni-processor real time embedded systems using regenerative energy. Task scheduling should account for the properties of the regenerative energy source which fluctuates, capacity of the energy storage as well as deadlines of the time critical tasks that characterize most of real time embedded systems. In this context, designing efficient scheduling strategies is significantly more complex compared to conventional real-time scheduling. In this paper we compare several scheduling heuristics with the optimal algorithm known as LSA (Lazy Scheduling Algorithm). We report results of an experiment study in terms of percentage of deadlines satisfied.

Summary (3 min read)

1. Introduction

  • Real-time embedded systems play a considerable role in their society, and they cover a spectrum from the very simple to the very complex.
  • Harvesting systems constructed to date extract power efficiently from the source.
  • Then, the crucial issue is to find scheduling mechanisms that can adapt the performance to the available energy profile.
  • Section 3 gives background materials around real-time scheduling and precisely describes the optimal uni-processor scheduler, LSA, under energy harvesting assumptions.

2. 1 Motivations

  • With the multitude of mobile devices that are used in all areas of social and commercial life it is increasingly important to design systems in an energy-efficient manner.
  • The true autonomy of such systems depends on their reliable and guaranteed operation for extended times without maintenance.
  • In addition to the very slow growth in their energy capacity, traditional batteries have a limit to the total practical energy density they can provide.
  • Economical deployment, embedded systems must gather energy from the environment around it, a technique known as energy harvesting or energy scavenging.
  • While some of these applications are marginal today, they will become commonplace one day.

2.2 History

  • A number of projects have used energy harvesting technologies to deliver sustainable power for autonomous sensors.
  • The solar panel is directly connected to its battery through a diode.
  • Prometheus has a super-capacitor as a primary buffer, a Li-Polymer battery, and a solar panel [2].
  • The common drawback of these first prototypes is that they do not target at real-time and quality of service requirements that characterize most of embedded applications.
  • Several prototype systems incorporating vibration energy harvesting have been developed too.

3.1 Background materials

  • Most of embedded applications require periodic activities that have to be cyclically executed at fixed rates and within special deadlines.
  • Schedulability analysis of periodic task sets can easily be performed both under fixed and dynamic priority assignments.
  • Nevertheless, they did not consider the rechargeability of the batteries.
  • EDF and RM scheduling have been extended to variable-voltage processors.
  • But solely applying these techniques has limitations in energy harvesting systems because they minimize CPU power, rather than they dynamically manage power according to the profiles of both available energy and processor workload.

3.2 An optimal scheduling algorithm

  • The first work that really makes adaptive power management for energy harvesting systems with real time constraints has been published in [9].
  • There, C. Moser et al. propose a real-time scheduling algorithm, called Lazy Scheduling Algorithm (LSA) that uses task postponement.
  • The electrical energy can be stored in the energy storage , whose capacity is precisely known.
  • So, in practice, the authors have to consider its computational overhead, i.e. the cost of its operation both in terms of time and energy consumption.

4. Description of scheduling heuristics

  • To evaluate the effectiveness of the LSA algorithm on energy saving and performance improvement, the authors developed a discrete-event simulation and compared LSA to several scheduling heuristics, all using the simple and easy to implement earliest deadline rule: ● Heuristic 1: EDt.
  • All the tasks are processed as soon as possible according to EDF until the battery is empty.
  • Then, the processor is put into sleep mode until the next release date i.e. the next instant corresponding to the arrival of another ready task.
  • These are typical greedy scheduling strategies which mainly differ each other in the way of managing energy lack situations.
  • In contrast, EDt will test dynamically energy availability before running tasks and does not greedily consumes energy.

5.1 Description of the simulator

  • In order to evaluate and compare the performance and the effectiveness of the scheduling heuristics to the optimal algorithm, the authors developed a discrete-event simulation in C/C++.
  • The simulator has been designed specifically for any periodic task set under energy constraints.
  • In their study, the authors consider the scheduling problem of periodic tasks.
  • The rechargeable power is variable with time in practice.
  • In other terms, 10/6 units of time are required to replenish the energy storage from the environmental power source.

5.2 Experiment 1: Varying the processing load

  • Figure 1 presents the ratios of satisfied deadlines for the heuristics and the optimal algorithm.
  • Figure 1 naturally shows that optimal LSA outperforms all other policies.
  • Velocities of degradations for the other heuristics are intermediate.
  • And for 100% processor utilization, LSA succeeds in satisfying about 60% task deadlines while the performance of EDd drops to less than 10%.
  • For every processor load, EDd provide the worst performance.

5.3 Experiment 2: Varying the battery capacity

  • The authors choose three different values for U, 0.2, 0.5 and 0.8 respectively representing low, medium and high system load.
  • When the battery capacity is less than 12, difference of performance between LSA and EDi is very large and as the battery capacity increases, the gaps are getting smaller.
  • For a battery capacity equal to 3, 100% deadlines are satisfied under LSA, Edi and EDu while less than 10% respectively 30% under EDd respectively EDt and EDc.
  • The authors see that if they make the load 2.5 times, they have to make the battery capacity more than 10 times to guarantee the same level of performance.
  • Figure 2 (c) reports the results for high processing loads (U=0.8).

6. Conclusion

  • Careful energy management is the key to providing the best possible performance in real-time harvesting systems.
  • The authors have implemented the policies and reported results showing that the optimal policy outperforms the heuristics that they examined.
  • Results were in terms of percentage of satisfied deadlines which is commonly used to measure the performance of real time systems.
  • This interesting issue needs more attention that will be in their immediate research plan.
  • The authors are measuring the impact of approximating energy availability on the effective performance of LSA and the actual gain of LSA if still existing.

Did you find this useful? Give us your feedback

Content maybe subject to copyright    Report

HAL Id: hal-00541127
https://hal.archives-ouvertes.fr/hal-00541127
Submitted on 29 Nov 2010
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entic research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diusion de documents
scientiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Performance Evaluation of Real-Time Scheduling
Heuristics for Energy Harvesting Systems
Maryline Chetto, Hui Zhang
To cite this version:
Maryline Chetto, Hui Zhang. Performance Evaluation of Real-Time Scheduling Heuristics for Energy
Harvesting Systems. The 2010 International Symposium on Energy-aware Computing and Networking
(EaCN-2010), Dec 2010, Hangzhou, China. pp.0. �hal-00541127�

Performance Evaluation of Real-Time Scheduling Heuristics
for Energy Harvesting Systems
Maryline CHETTO and Hui ZHANG
IRCCyN University of Nantes
1 Rue de la Noé, F-44321 Nantes FRANCE
Maryline.chetto@univ-nantes.fr
Hui.zhang@univ-nantes.fr
Abstract
Energy constrained systems can increase their usable
lifetimes by extracting energy from their environment.
This is known as energy harvesting. This paper
investigates scheduling issues in uni-processor real
time embedded systems using regenerative energy.
Task scheduling should account for the properties of
the regenerative energy source which fluctuates,
capacity of the energy storage as well as deadlines of
the time critical tasks that characterize most of real
time embedded systems. In this context, designing
efficient scheduling strategies is significantly more
complex compared to conventional real-time
scheduling. In this paper we compare several
scheduling heuristics with the optimal algorithm
known as LSA (Lazy Scheduling Algorithm). We report
results of an experiment study in terms of percentage
of deadlines satisfied.
1. Introduction
Real-time embedded systems play a considerable
role in our society, and they cover a spectrum from the
very simple to the very complex. Examples of current
systems include the control of domestic appliances like
washing machines and televisions, the control of
automobile engines, telecommunication systems,
military command and control systems, industrial
process control, flight control systems, and space
shuttle and aircraft avionics... For example, a system
that monitors temperature in a nuclear power plant
would require that the readings be reported to a base
station within enough time for a proper response to be
made to a rapid increase in the temperature.
Harvesting energy in surrounding environment to
power embedded systems for the lifespan appears
nowadays to be the alternative to conventional
batteries. Harvesting systems constructed to date
extract power efficiently from the source. However,
they do not use it adequately under real-time running
conditions. As a result, they need a much larger
harvester (e.g. solar panel) than necessary to yield the
same level of power as a more efficient one, or they
rely on a larger, more expensive, higher capacity
battery than needed in order to sustain extended
operation.
In this new context, the main problem to solve
comes from the instantaneous power level that tends to
vary over a wide. The autonomous nature of operation
makes it imperative that the system learns its own
energy environment and adapts its power consumption
accordingly. Goal of this adaptation is to maximize the
utility of the application in a long-term perspective.
The resulting mode of operation is sometimes called
energy neutral operation.
Then, the crucial issue is to find scheduling
mechanisms that can adapt the performance to the
available energy profile. Up to now, when designing a
real-time embedded system, the first concern has been
usually time, leaving energy efficiency as a hopeful
consequence of empiric decisions. Now, the primary
concern is that power from solar panels or other free
sources that cannot be stored (or stored with limited
capacity) should be fully consumed greedily, or else
this energy will be wasted.

In a real-time environment where tasks have to meet
deadlines and execute periodically, energy harvesting
and task scheduling are strongly dependent since they
have to jointly handle timing constraints and
variability of available energy.
In that paper, we propose five scheduling heuristics,
easy to implement with limited overhead. Our
objective is to compare their performance to an
optimal scheduler, namely LSA with high overhead.
The main question is to appreciate precisely the
relative performance of all these schedulers thanks to
simulation. All of them are based on the famous
Earliest Deadline First rule. We report results of an
experiment study in terms of percentage of satisfied
deadlines and wasted energy. The quantitative analysis
is achieved without considering computation
overheads.
The remainder of this paper is organized as follows.
In section 2, we describe the main issues in energy
harvesting and briefly describe projects that involve
energy harvesting. Section 3 gives background
materials around real-time scheduling and precisely
describes the optimal uni-processor scheduler, LSA,
under energy harvesting assumptions. In section 4, we
present five scheduling heuristics and we propose to
compare them to LSA. Section 5 reports results of a
simulation study and enables us to bring to light that
some heuristics may have similar performance to LSA
without incurring high overhead. We conclude in
section 6 with a brief discussion on ongoing research.
2. Energy harvesting
2. 1 Motivations
With the multitude of mobile devices that are used in
all areas of social and commercial life it is increasingly
important to design systems in an energy-efficient
manner. These autonomous systems (including sensor-
actuator networks) are being envisioned to carry out
complex task sets under real-time requirements
without human intervention. However, they require
power in order to operate, and if power outages occur,
critical data may be lost. The true autonomy of such
systems depends on their reliable and guaranteed
operation for extended times without maintenance.
Most prior wireless monitoring systems in last
decades have relied on continuous power supplied by
batteries such as lithium-ion cells. Their disadvantage
is that they become depleted, must be periodically
replaced or recharged and consequently place hard
restrictions on products' usability, lifetime, and cost of
ownership. Moreover, while processing power roughly
doubles ever two years, battery technology advances at
a much more sluggish pace (battery capacity has
doubled every 10 years). In addition to the very slow
growth in their energy capacity, traditional batteries
have a limit to the total practical energy density they
can provide.
Even if it is possible to increase their energy density
by tenfold within a few years, we must still consider
practical safety concerns. First, given improper use,
batteries with extremely high energy densities can
become dangerous, explosive devices. Second, in
many embedded applications, battery replacement is
impractical or has high labour costs associated with
maintenance.
Besides, batteries suffer from self-discharge,
temperature, and other environmental effects that work
to bound their usable lifetime, even in the case of zero
use. Consequently for long-term, economical
deployment, embedded systems must gather energy
from the environment around it, a technique known as
energy harvesting or energy scavenging.
It concerns as well the high technology sectors as the
general public products in which wireless embedded
systems are used in a variety of applications, such as
environmental applications (forest fire and flood
detection, monitoring of drinking water and level air
pollution), military applications (battlefield
surveillance, reconnaissance of enemy forces), health
applications (tele-monitoring of human physiological
data, tracking and monitoring of doctors and patients),
home applications (intrusion detection), or commercial
applications (monitoring of product quality, climate
control in large buildings). While some of these
applications are marginal today, they will become
commonplace one day. Devices with maintenance-free
life of hundreds of years can now be envisaged if we
provide them with efficient strategies for harvesting,
storing and managing environmental energy. The
current perspectives of this market are thus very big
and promising.
2.2 History
A number of projects have used energy harvesting
technologies to deliver sustainable power for
autonomous sensors. Photovoltaic energy harvesting is
by far the most prevalent form of technology used in
current projects in part due to the plentiful supply of
light in many deployment settings, and the low cost of
photovoltaic modules. Nodes conventionally store
electrical energy in super-capacitors or batteries to
achieve operation. Heliomote has a solar panel and two
AA type Ni MH batteries [1]. The solar panel is
directly connected to its battery through a diode. Even
though ample power may be available on the solar
panel, a wireless sensor node can still draw current
from the battery. Prometheus has a super-capacitor as

a primary buffer, a Li-Polymer battery, and a solar
panel [2]. The solar panel first charges the super-
capacitor, from which the system draws current when
enough power is available on the solar panel. The
system draws current from the battery only when the
charge level of the primary buffer is less than a certain
threshold, and it seldom draws power from the battery.
Heliomote and Prometheus have permitted to show
that systems may operate perpetually through
scavenging solar energy. However, the common
drawback of these first prototypes is that they do not
target at real-time and quality of service requirements
that characterize most of embedded applications.
Several prototype systems incorporating vibration
energy harvesting have been developed too. For
example, the S5NAP uses a commercially-available
electromagnetic vibration energy harvester to power
an accelerometer based condition monitoring system.
In this system, energy harvested from vibrations is
buffered in super-capacitors to permit nodes to draw
large bursts of power during radio transmissions and
sensing operations [3].
Another project named ShiMmer uses piezoelectric
transducers to evaluate a portion of a structure (i.e. a
bridge) to determine if damage exists. It relies on a
wireless platform that combines active sensing and
localized processing with energy harvesting to provide
long-lived structural health monitoring. One of the
objectives of ShiMmer project is to create a robust and
flexible software controller that can manage both the
energy and the task execution [4].
3. Scheduling with energy constraints
3.1 Background materials
Most of embedded applications require periodic
activities that have to be cyclically executed at fixed
rates and within special deadlines. Typically, each
periodic instance is assigned a relative deadline equal
to the task period and is treated as a hard job. Thus, a
periodic task is executed only if all its instances are
guaranteed to complete within their deadlines.
Schedulability analysis of periodic task sets can easily
be performed both under fixed and dynamic priority
assignments. In particular, a lot of work has been done
for the Rate Monotonic (RM) and the Earliest
Deadline First (EDF) algorithms [5]. Schedulability
analysis has also been extended for the case in which
tasks use shared resources or run in the presence of
aperiodic activities, under fixed priority scheduling
and in dynamic priority systems as well [6] [7].
While EDF (dynamic priority depending on urgency)
and RM (fixed priority depending on period) can
support sophisticated task set characteristics such as
deadlines, precedence constraints, shared resources,
jitter, etc., they are all open loop scheduling
algorithms. Open loop refers to the fact that once
schedules are created they are not "adjusted" based on
continuous feedback. Systems with open-loop
schedulers are usually designed based on worst-case
parameters. Such an approach can result in a highly
underutilized system based on extremely pessimistic
estimation of workload (or energy). While open-loop
scheduling algorithms can perform well when the
workload and the processing performance are
accurately modelled, they perform poorly in
unpredictable dynamic systems including regenerative
energy dependent ones.
Only in the past decade, researchers started to
address power and scheduling issues with the objective
of either minimizing power usage under timing
constraints or maximizing the system performance
under the energy constraints. Nevertheless, they did
not consider the rechargeability of the batteries. For
example, EDF and RM scheduling have been extended
to variable-voltage processors. The idea is to save
power by slowing down the processor just enough to
meet the deadlines [8]. But solely applying these
techniques has limitations in energy harvesting
systems because they minimize CPU power, rather
than they dynamically manage power according to the
profiles of both available energy and processor
workload.
The performance of a practical energy harvesting
real-time system is measured by the deadline miss rate
and heavily depends upon the stored energy and the
energy harvested from the environment.
Unfortunately, the scavenging power is time-varying
and thus very unstable. Therefore, the accurate
modelling for energy source plays a key role in
designing a good policy to schedule the tasks and
reduce the deadline miss rate.
3.2 An optimal scheduling algorithm
The first work that really makes adaptive power
management for energy harvesting systems with real
time constraints has been published in [9]. There, C.
Moser et al. propose a real-time scheduling algorithm,
called Lazy Scheduling Algorithm (LSA) that uses
task postponement. Algorithm LSA is energy-
clairvoyant, i.e., the generated energy in the future is
known. Taking into account available time as well as
processable energy, an optimal task ordering can be
determined based on the prediction of the available
energy in the future.
This work deals with a mono-processor architecture
that draws the energy from storage and uses it to

process tasks (periodic or non periodic) with arrival
time, deadline, and worst case execution time. The
worst case execution time represents the maximum
energy demand of the task. The arrival time of the task
is not known beforehand. The deadline as well as the
worst case execution time of the task is unknown
before it is released. However, as long as the task is
released, all these parameters are determined. They
assume that tasks are preemptable and execute
according to the earliest deadline first policy.
At any time, the energy source module harvests the
energy from its ambient environment and then
converts it into electrical energy. The electrical energy
can be stored in the energy storage (battery), whose
capacity is precisely known. The stored energy is
assumed to be known at the system level at any time
and is no more than the storage capacity. It is assumed
that the energy storage is ideal and the battery can be
recharged up to its capacity. Likewise, it can also be
completely discharged to as less as zero. If the stored
energy reaches the capacity, the incoming harvested
energy overflows the storage and is discarded.
According to LSA, the processor executes all tasks at
full power when the battery is full time, and the system
starts executing a task if the task is ready and has the
earliest deadline among all ready tasks and the system
is able to keep on running at the maximum power until
the deadline of the task.
Contrary to greedy scheduling algorithms, LSA
hesitates to power tasks until it is necessary to respect
timing constraints. In that sense, tasks are executed
neither as soon as possible nor as late as possible. In
this paper, the authors also discuss an admittance test
that decides, whether a set of real-time tasks can be
scheduled without violating deadlines. Another crucial
question which has been solved is how to dimension
the capacity of the battery that ensures continuous
operation. The simulation study demonstrates that
achievable capacity savings between 20% and 45% are
obtained comparing the classical Earliest Deadline
First algorithm. However, all theses measurements
ignore on line computational costs.
While optimal in the case of a single speed
processor, LSA algorithm has the following
drawbacks:
The consumption power of the task is assumed to
be characterized by some value. This implies that for
every task, its total energy consumption is directly
connected to its execution time through the constant
power of the processing device. However, in practice,
the total energy which can be consumed by a task has
no correlation with the worst case execution time.
Renewable energy sources must be accurately
modelled, otherwise the performance of LSA will be
degraded.
Scheduler LSA requires a lot of mathematical
computations to be performed on-line. So, in practice,
we have to consider its computational overhead, i.e.
the cost of its operation both in terms of time and
energy consumption.
4. Description of scheduling heuristics
To evaluate the effectiveness of the LSA algorithm
on energy saving and performance improvement, we
developed a discrete-event simulation and compared
LSA to several scheduling heuristics, all using the
simple and easy to implement earliest deadline rule:
Heuristic 1: EDt. Before starting the execution of
the highest priority task which is ready, a test is
performed to compare the energy level of the battery
to the total energy required by the task for its
execution. If the energy available in battery is
sufficient, the task is authorized to execute. Otherwise,
the processor is put into sleep mode until the battery
contains enough energy to run it. According to this
scheduler, we never start execution of a task if there is
no sufficient energy to execute it totally.
Heuristic 2: EDi. All the tasks are processed as
soon as possible according to EDF until the battery is
empty. Then, the processor is put into sleep mode until
the next release date i.e. the next instant corresponding
to the arrival of another ready task. During that period,
the battery will replenish and necessarily, the
processor will be active at that instant for executing
the highest priority task.
Heuristic 3: EDd. As previously, all tasks execute
as soon as possible according to EDF until there is no
more energy available in the battery. Then, all ready
tasks are discarded and the processor is put into sleep
mode until the next release time.
Heuristic 4: EDu. Compared to EDi, EDu is
similar but lets the processor in sleep mode just during
one time unit whenever the battery is empty.
Heuristic 5: EDc. Compared to EDd, EDc is
similar but just deletes the current active task instead
of discarding all the tasks waiting for execution. The
processor is put into sleep mode until the arrival of a
new task even if the list of ready tasks may not be
empty.
Edi, EDd, EDu and EDc execute tasks as soon as
possible i.e. as long as the battery contains at least one
unit of energy. These are typical greedy scheduling
strategies which mainly differ each other in the way of
managing energy lack situations. In contrast, EDt will
test dynamically energy availability before running
tasks and does not greedily consumes energy. Let us
note that for every heuristic, as soon as a deadline is
missed, the corresponding task is aborted for

Citations
More filters
Journal ArticleDOI
TL;DR: This work addresses the problem of task scheduling in processors located in sensor nodes powered by EH sources, and proposes a new improved LSA approach, namely energy-aware LSA, which is applied in order to reduce the LSA computational complexity and thus maximizing the amount of energy available for task execution.
Abstract: The main problem in dealing with energy-harvesting (EH) sensor nodes is represented by the scarcity and non-stationarity of powering, due to the nature of the renewable energy sources. In this work, the authors address the problem of task scheduling in processors located in sensor nodes powered by EH sources. Some interesting solutions have appeared in the literature in the recent past, as the lazy scheduling algorithm (LSA), which represents a performing mix of scheduling effectiveness and ease of implementation. With the aim of achieving a more efficient and conservative management of energy resources, a new improved LSA solution is here proposed. Indeed, the automatic ability of foreseeing at run-time the task energy starving (i.e. the impossibility of finalizing a task due to the lack of power) is integrated within the original LSA approach. Moreover, some modifications have been applied in order to reduce the LSA computational complexity and thus maximizing the amount of energy available for task execution. The resulting technique, namely energy-aware LSA, has then been tested in comparison with the original one, and a relevant performance improvement has been registered both in terms of number of executable tasks and in terms of required computational burden.

22 citations

Journal ArticleDOI
TL;DR: This article proposes an indoor test methodology that does not rely on solar simulators and has its basis in astronomy and photovoltaic cell design, and presents a generic design for a test apparatus that can be used in carrying out the test methodology.
Abstract: Repeatable and accurate tests are important when designing hardware and algorithms for solar-powered wireless sensor networks (WSNs). Since no two days are exactly alike with regard to energy harvesting, tests must be carried out indoors. Solar simulators are traditionally used in replicating the effects of sunlight indoors; however, solar simulators are expensive, have lighting elements that have short lifetimes, and are usually not designed to carry out the types of tests that hardware and algorithm designers require. As a result, hardware and algorithm designers use tests that are inaccurate and not repeatable (both for others and also for the designers themselves). In this article, we propose an indoor test methodology that does not rely on solar simulators. The test methodology has its basis in astronomy and photovoltaic cell design. We present a generic design for a test apparatus that can be used in carrying out the test methodology. We also present a specific design that we use in implementing an actual test apparatus. We test the efficacy of our test apparatus and, to demonstrate the usefulness of the test methodology, perform experiments akin to those required in projects involving solar-powered WSNs. Results of the said tests and experiments demonstrate that the test methodology is an invaluable tool for hardware and algorithm designers working with solar-powered WSNs.

13 citations

Proceedings ArticleDOI
01 Oct 2013
TL;DR: This paper combines the use of DVS and energy harvesting with the capabilities of a feedback scheduler to minimize the consumed energy, as well as to take the charging model into account.
Abstract: The use of environmental energy such as solar energy has recently emerged as an option to increase the operating time of embedded systems (e.g. wireless sensors). It consists in converting ambient energy into electricity to power and lengthen battery life. Dynamic voltage scaling (DVS) is one of the most effective techniques for reducing energy consumption in embedded and real-time systems. However, traditional DVS algorithms have inherent limitations on their capability in energy saving since they rarely take into account the actual application requirements and often exploit fixed timing constraints of the real-time tasks. Some authors used feedback scheduling techniques in order to minimize the consumed energy by observing the actual usage of resources in the system. This paper combines the use of DVS and energy harvesting with the capabilities of a feedback scheduler. Our goal is to minimize the consumed energy, as well as to take the charging model into account.

6 citations

Journal ArticleDOI
TL;DR: This paper compares several battery thresholds knowing that the processor performs the tasks at maximum speed when the battery storage is over the thresholds, and evaluation of this new real-time feedback scheduler shows experimentally a good compromise between the available energy and the processor workload.

4 citations


Cites background from "Performance Evaluation of Real-Time..."

  • ...However, it needs an exponential complexity for periodic task sets, as well as a high complexity for some kinds of complex energy harvesting curves [18]....

    [...]

Book ChapterDOI
11 Jul 2012
TL;DR: The authors' objective consists in employing a conservative scheduling paradigm in order to achieve a more efficient management of energy resources and to prove such a claim, the recently advanced Lazy Scheduling Algorithm has been taken as reference and integrated with the automatic ability of foreseeing at runtime the task energy starving.
Abstract: One of the most challenging issues for nowadays Wireless Sensor Networks (WSNs) is represented by the capability of self-powering the network sensor nodes by means of suitable Energy Harvesting (EH) techniques. However, the nature of such energy captured from the environment is often irregular and unpredictable and therefore some intelligence is required to efficiently use it for information processing at the sensor level. In particular in this work the authors address the problem of task scheduling in processors located in WSN nodes powered by EH sources. The authors' objective consists in employing a conservative scheduling paradigm in order to achieve a more efficient management of energy resources. To prove such a claim, the recently advanced Lazy Scheduling Algorithm (LSA) has been taken as reference and integrated with the automatic ability of foreseeing at runtime the task energy starving, i.e. the impossibility of finalizing a task due to the lack of power. The resulting technique, namely Energy Aware Lazy Scheduling Algorithm (EA-LSA), has then been tested in comparison with the original one and a relevant performance improvement has been registered in terms of number of executable tasks.

4 citations

References
More filters
Journal ArticleDOI
TL;DR: The problem of multiprogram scheduling on a single processor is studied from the viewpoint of the characteristics peculiar to the program functions that need guaranteed service and it is shown that an optimum fixed priority scheduler possesses an upper bound to processor utilization.
Abstract: The problem of multiprogram scheduling on a single processor is studied from the viewpoint of the characteristics peculiar to the program functions that need guaranteed service. It is shown that an optimum fixed priority scheduler possesses an upper bound to processor utilization which may be as low as 70 percent for large task sets. It is also shown that full processor utilization can be achieved by dynamically assigning priorities on the basis of their current deadlines. A combination of these two scheduling techniques is also discussed.

7,067 citations


"Performance Evaluation of Real-Time..." refers background in this paper

  • ...In particular, a lot of work has been done for the Rate Monotonic (RM) and the Earliest Deadline First (EDF) algorithms [ 5 ]....

    [...]

Book
03 Jan 1989
TL;DR: In this paper, the problem of multiprogram scheduling on a single processor is studied from the viewpoint of the characteristics peculiar to the program functions that need guaranteed service, and it is shown that an optimum fixed priority scheduler possesses an upper bound to processor utilization which may be as low as 70 percent for large task sets.
Abstract: The problem of multiprogram scheduling on a single processor is studied from the viewpoint of the characteristics peculiar to the program functions that need guaranteed service. It is shown that an optimum fixed priority scheduler possesses an upper bound to processor utilization which may be as low as 70 percent for large task sets. It is also shown that full processor utilization can be achieved by dynamically assigning priorities on the basis of their current deadlines. A combination of these two scheduling techniques is also discussed.

5,397 citations

Journal Article
TL;DR: In this article, the authors describe key issues and tradeoffs which arise in the design of solar energy harvesting, wireless embedded systems and present the design, implementation, and performance evaluation of Heliomote, their prototype that addresses several of these issues.
Abstract: Sustainable operation of battery powered wireless embedded systems (such as sensor nodes) is a key challenge, and considerable research effort has been devoted to energy optimization of such systems. Environmental energy harvesting, in particular solar based, has emerged as a viable technique to supplement battery supplies. However, designing an efficient solar harvesting system to realize the potential benefits of energy harvesting requires an in-depth understanding of several factors. For example, solar energy supply is highly time varying and may not always be sufficient to power the embedded system. Harvesting components, such as solar panels, and energy storage elements, such as batteries or ultracapacitors, have different voltage-current characteristics, which must be matched to each other as well as the energy requirements of the system to maximize harvesting efficiency. Further, battery nonidealities, such as self-discharge and round trip efficiency, directly affect energy usage and storage decisions. The ability of the system to modulate its power consumption by selectively deactivating its sub-components also impacts the overall power management architecture. This paper describes key issues and tradeoffs which arise in the design of solar energy harvesting, wireless embedded systems and presents the design, implementation, and performance evaluation of Heliomote, our prototype that addresses several of these issues. Experimental results demonstrate that Heliomote, which behaves as a plug-in to the Berkeley/Crossbow motes and autonomously manages energy harvesting and storage, enables near-perpetual, harvesting aware operation of the sensor node.

1,063 citations

Proceedings ArticleDOI
24 Apr 2005
TL;DR: Prometheus as discussed by the authors is a two-stage energy storage system consisting of supercapacitors (primary buffer) and a lithium rechargeable battery (secondary buffer), which can operate for 43 years under 1% load and 4 years under 10% load.
Abstract: Environmental energy is an attractive power source for low power wireless sensor networks. We present Prometheus, a system that intelligently manages energy transfer for perpetual operation without human intervention or servicing. Combining positive attributes of different energy storage elements and leveraging the intelligence of the microprocessor, we introduce an efficient multi-stage energy transfer system that reduces the common limitations of single energy storage systems to achieve near perpetual operation. We present our design choices, tradeoffs, circuit evaluations, performance analysis, and models. We discuss the relationships between system components and identify optimal hardware choices to meet an application's needs. Finally we present our implementation of a real system that uses solar energy to power Berkeley's Telos Mote. Our analysis predicts the system will operate for 43 years under 1% load, 4 years under 10% load, and 1 year under 100% load. Our implementation uses a two stage storage system consisting of supercapacitors (primary buffer) and a lithium rechargeable battery (secondary buffer). The mote has full knowledge of power levels and intelligently manages energy transfer to maximize lifetime.

803 citations

Proceedings ArticleDOI
24 Apr 2005
TL;DR: Experimental results demonstrate that Heliomote, which behaves as a plug-in to the Berkeley/Crossbow motes and autonomously manages energy harvesting and storage, enables near-perpetual, harvesting aware operation of the sensor node.
Abstract: Sustainable operation of battery powered wireless embedded systems (such as sensor nodes) is a key challenge, and considerable research effort has been devoted to energy optimization of such systems. Environmental energy harvesting, in particular solar based, has emerged as a viable technique to supplement battery supplies. However, designing an efficient solar harvesting system to realize the potential benefits of energy harvesting requires an in-depth understanding of several factors. For example, solar energy supply is highly time varying and may not always be sufficient to power the embedded system. Harvesting components, such as solar panels, and energy storage elements, such as batteries or ultracapacitors, have different voltage-current characteristics, which must be matched to each other as well as the energy requirements of the system to maximize harvesting efficiency. Further, battery non-idealities, such as self-discharge and round trip efficiency, directly affect energy usage and storage decisions. The ability of the system to modulate its power consumption by selectively deactivating its sub-components also impacts the overall power management architecture. This paper describes key issues and tradeoffs which arise in the design of solar energy harvesting, wireless embedded systems and presents the design, implementation, and performance evaluation of Heliomote, our prototype that addresses several of these issues. Experimental results demonstrate that Heliomote, which behaves as a plug-in to the Berkeley/Crossbow motes and autonomously manages energy harvesting and storage, enables near-perpetual, harvesting aware operation of the sensor node.

506 citations

Frequently Asked Questions (1)
Q1. What are the contributions mentioned in the paper "Performance evaluation of real-time scheduling heuristics for energy harvesting systems" ?

This paper investigates scheduling issues in uni-processor real time embedded systems using regenerative energy. In this paper the authors compare several scheduling heuristics with the optimal algorithm known as LSA ( Lazy Scheduling Algorithm ). The authors report results of an experiment study in terms of percentage of deadlines satisfied.