Energy Efficiency of the Swarm-Capable Micro AUV SEMBIO

Abstract

The power module is one of the most important components of an AUV's electrical design. The AUV's propulsion , sensors, and processing are usually powered by batteries. However, batteries cannot provide a great deal of energy, which influences the AUV's range and mission objectives, and, implicitly , its level of autonomy. For small and micro AUV, longer endurance is a key aim and a desirable feature for practical applications. In addition, returning the robot to its home-base and recharging the AUV to complete the task would be expensive and time-consuming. This paper presents the SEMBIO AUV that was constructed to reduce energy consumption using a streamlined body design. Although the enhancement of energy sources and storage performance is beyond the scope of this work, attention must be paid not only to the performance of the on-board energy source, but also to the amount of energy that is required overall. The experimental results of reducing the amount of energy consumption by SEMBIO during operation by a solar and energy management system (SEMS) are conducted.

Full text

Energy Efficiency of the Swarm-Capable Micro AUV SEMBIO
Ammar Amory and Erik Maehle
University of Luebeck - Institute of Computer Engineering
Luebeck, Germany
Email: {amory, maehle}@iti.uni-luebeck.de
Abstract—The power module is one of the most important
components of an AUV’s electrical design. The AUV’s propul-
sion, sensors, and processing are usually powered by batteries.
However, batteries cannot provide a great deal of energy, which
influences the AUV’s range and mission objectives, and, implic-
itly, its level of autonomy.
For small and micro AUV, longer endurance is a key aim
and a desirable feature for practical applications. In addition,
returning the robot to its home-base and recharging the AUV to
complete the task would be expensive and time-consuming.
This paper presents the SEMBIO AUV that was constructed
to reduce energy consumption using a streamlined body de-
sign. Although the enhancement of energy sources and storage
performance is beyond the scope of this work, attention must
be paid not only to the performance of the on-board energy
source, but also to the amount of energy that is required overall.
The experimental results of reducing the amount of energy
consumption by SEMBIO during operation by a solar and energy
management system (SEMS) are conducted.
Index Terms—autonomous underwater vehicle, swarm, energy
consumption, vehicle design.
I. INTRODUCTION
Unlike tethered ROVs that can receive their power from
the mission vessel, untethered vehicles are restricted in their
duration of operation by their on-board power system. For the
majority of existing AUVs, the system is based on batteries
that provide a limited amount of energy [1]. The hardware
or software structure of underwater robots was improved in a
number of projects as a way of the extending mission duration,
as in the case of the SAUV [2].
For autonomous vehicles intended for long-duration mis-
sions, the essential design components are the on-board energy
source and storage [3]. Autonomous vehicles will be able to
endure longer if the amount of on-board energy per unit weight
is high, irrespective of additional factors.
The energy resource constitutes an important factor in the
context of swarm robotics. The purpose of the swarm robotics
strategy is to establish coordination among multiple robots that
feature sensing, actuation, processing, and cognition abilities
to cooperate on the achievement of a common task [4], [5], [6],
[7]. Therefore, AUVs are the main ingredient or the key com-
ponent to creating underwater swarms. The SEMBIO AUV (as
shown in Fig 1) is a vehicle with high maneuverability that is
designed and developed to enable the use of marine robots in
swarm applications. The main uses of SEMBIO AUVs are to
explore and monitor the coastal environment.
Energy management is so important in the case of AUVs
because they are battery-powered, meaning that the energy
amount available to them is limited, which has implications
Figure 1. The SEMBIO AUV developed by the University of Luebeck. The
design was inspired by the hydrodynamic features of a guitarfish leading to a
highly energy-efficient vehicle using cost-effective hardware and controllers
such as simple electronic speed controllers [8]
.
for the duration of operation and the types of missions that can
be undertaken. To achieve optimal regulation of power usage,
numerous tradeoffs between data processing, active sensing,
navigation, and communication are necessary. Therefore, given
that the management and control of AUV energy resources
is of such significance, a system capable of measuring the
power used by every AUV component and reducing energy
consumption whenever possible to prolong the mission time
has been developed.
The features of the SEMBIO power system are discussed
in the following section, with particular emphasis on the
management of the overall on-board power based on an energy
management system capable of monitoring all data regarding
energy consumption and the mission time. Furthermore, to en-
able vehicle localization and retrieval in emergency situations,
and to provide a backup for the primary energy system, a solar
system was constructed.
II. SEMBIO AUV
The SEMBIO AUV, illustrated in Fig 1, developed by
the Institute of Computer Engineering of the University of
Luebeck is such a micro AUV designed as a highly energy
efficient vehicle, using an optimized hydrodynamic design
and cost-effective hardware [9]. The design was inspired by
the hydrodynamic features of a guitarfish. The SEMBIO AUV
has some characteristics making it very suitable for smooth
and flexible motions on the one hand and highly appropriate
for a variety of AUV swarm applications on the other hand.
It utilizes conventional thrusters including simple electronic
speed controllers (ESCs), mainly used for quadrocopters and

Figure 2. Illustration of the fluid flow during the CFD analysis of SEMBIO.
The streamline features of the robot’s hull are validated by CFD [8]
.
other robots in the field of model making.
The robustness of the design was improved by extensive and
detailed finite element analysis (FEA). Hence, the main aspect
of the design is the robot’s form, based on the body of the
guitar fish, which minimizes water resistance and generates
a smooth fluid flow. The hull design and its hydrodynamic
resistance is one of the most important factors directly af-
fecting power requirements and the maneuverability of the
vehicle. The minimizing of the hydrodynamic resistance of
AUVs leads to high energy efficiency. As a consequence, this
increases the speed of movement in water and reduces energy
consumption of the propulsion system. The streamline features
were proven by Computational Fluid Analysis (CFD) [8]
illustrated in Fig 2. Improvement and optimization of the
3D shape model were achieved via CFD simulations, which
were useful for the assessment of the hydrodynamic properties
of the proposed vehicle and also provided guidelines for
enhancing its design. As a result of the CFD analysis, pressure
and velocity distributions over the AUV bodies were obtained.
The CFD results indicated that the streamlined SEMBIO hull
has a small drag coefficient, which leads SEMBIO to consume
less energy or showing a higher energy efficiency thanks to
its geometric shape as shown in Fig 2.
A. Power System Design
The components of this system include a primary battery
powering the electronic components, a power switch high-
current relay for switching between the charger and the battery,
a power switch linking the battery to the external charger
via a tether, a magnetic switch, and a DC-DC converter for
power conversion into 11.1V, 5V, and 3.3V, according to
the requirements of the various electronic components. To
enable the batteries to be charged within the hull, a power
switch board was designed. A pair of diodes and relays were
included in the developed circuit board, which also had an
extension cable linking it to the external connector to ensure
the availability of a charger or an external source. Four of
the eight pin positions on a circular connector positioned on
SEMBIO’s tail end were used by the charger, while the other
pin positions facilitated transferring the data between the robot
and a computer. Due to the ease of plugging in and charging
the vehicle prior to a mission or work in the preparatory stage
in extreme conditions, such a feature is useful for the work
between missions.
B. SEMBIO’s Batteries
SEMBIO was equipped with two lithium polymer (Li-
Po) batteries, namely, the primary battery and the emergency
battery, which are used respectively in regular and emergency
conditions. For the application, the emergency battery could
be more significant if it could be charged during the mission.
SEMBIO has an overall power capacity of around 120 Wh (as
shown in Table I).
Table I
THE BATT ERY S PEC IFI CATIO N OF SEMBIO’S P OWE R SYS TE M
Battery type Voltage Capacity Weight Specific energy
Primary 11.1 V (3S) 10 Ah (111 Wh) 720 g 155 Wh/kg
Emergency 3.7 V (1S) 2.7 Ah (10 Wh) 50 g 185 Wh/kg
C. Solar Energy
SEMBIO was equipped with a solar system to prolong
mission duration and to enable vehicle retrieval in the event of
a power system error or energy exhaustion. The flexible solar
panel was chosen for use in the solar power system, as it is
capable of supplying around 6 V DC in direct sunlight at up
to 1 W or about 160 mA. Although it presents cost efficiency,
ease of maintenance, and ease of use, unfortunately this solar
panel is not as efficient as other alternatives.
III. SO LA R AN D EN ER GY M ANAGEMENT SYSTEM (SEMS)
Our previously paper published in 2016 illustrated the
preliminary results of a solar and energy management sys-
tem (SEMS) [9]. The solar and energy management system
(SEMS) was developed to ensure that the on-board energy
source was used effectively by keeping track of the energy
consumption of each component in the electrical system. The
creation of SEMS was justified by the need to control AUV
energy resources to reduce power usage and thus enhance
battery life, thereby prolonging mission endurance or time. An
additional objective of SEMS is to minimize the complexity
of AUV operational procedures for researchers and other users
to make it easier to use such vehicles for different missions.
A. SEMS Concept
Unlike the traditional approach of transmitting a mission to
a robot, which involves the functions of every robot component
throughout the mission, SEMS was designed to divide a
mission into tasks and subtasks. The subtask is the lowest
level of the mission structure that controls and measures the
energy consumption of every component of the AUV. For
example, the subtasks monitor the power usage of cameras,

acoustic communication, pressure sensors, thrusters, and pro-
cessing units. The major constituents of SEMS are SEMS-
Communication, SEMS-Embedded-Board, SEMS-Solar, and
SEMS-Mission-Planner.
The SEMS-Embedded-Board is installed in a shield of
Arduino Mega to lead the energy management processes on-
board the vehicle. Fig 3 illustrate the SEMS-Embedded-Board
prototype and its components. The vehicle has been equipped
with two main processors Arduino Mega and Raspberry Pi2.
Pressure and Temperature sensors are always required in un-
derwater robots. To facilitate the communication between the
energy management system and the GUI mission on PC, the
XBee [10] (ZigBee/IEEE 802.15.4 compliant) was selected.
For the system used in SEMBIO, the PXFmini [11] with
the MPU9250 inertial measurement unit (IMU) was chosen,
comprising a 3-axis accelerometer, a 3-axis magnetometer, and
a 3-axis gyro. The uBlox Neo-M8N GPS module with an
HMC5883L digital compass kit was chosen that delivers and
provides location and time information outdoors in all weather
conditions. SEMBIO was equipped with two cameras. One
camera was a forward-facing camera (RPi camera) that was
positioned at the front of the hull, while the other camera (Pixy
camera) was a downward-facing camera that was positioned at
the bottom of the vehicle’s hull. The purpose of both of these
cameras was to support the vision system of the vehicle.
XBee
Module
MicroSD
Card
ARM
Processor
Pixy
Camera
Wireless
Programming
Power
Supply
IMU
Sensor
GPS
Module
Camera
Light
Pressure
Sensor
ESCs and
Thrusters
Compass
Measurement Power Consumption Units (INA219)
Switching Units (SSRs)
Embedded Controller
Figure 3. Diagram showing the functional components of the SEMS-
Embedded-Board
A system is assumed that assists practitioners in building
AUV components, and in defining missions, tasks, and sub-
tasks (Fig 4). In this work, thruster unit power consump-
tion measurement was undertaken more deeply by SEMS.
It focuses on how and where consumption occurred and
how the measurement results could be employed to improve
performance while minimizing consumption.
IV. MEASUREMENT OF COMPONENT POWER
CONSUMPTION
The overall energy requirement or consumption of a mission
can be approximated through measuring the amount of energy
consumed by every component. SEMS evaluation can also
be conducted based on the development of a database of
component energy consumption. SEMS tries to use a high
range of energy efficiency since certain components (e.g.,
thrusters) possess this range.
Start
Task
(active and run all elements)
End
Task
Start
Subtask 1
End
Subtask 2
Subtask N
Task1
Traditional method SEMS method
Mission
Start
Subtask 1
End
Subtask 2
Subtask N
Task 2
Start
Subtask 1
End
Subtask 2
Subtask N
Task N
Mission
Figure 4. The different structures of a mission’s execution
SEMS itself is used for the measurement of power, be-
ginning by connecting a certain component to the equivalent
pin on the SEMS-Embedded-Board which is plugged into
Arduino Mega, and possesses 32 addressed pins for SEMS.
Subsequently, SEMS sends 0 to all components to switch them
off before employing the INA219 current sensor to measure
their current consumption. An independent current sensor is
included in the input of certain pins. Many current sensors
are incorporated in the SEMS-Embedded-Board for linking
to components such as cameras and Raspberry Pi (RPi), the
current usage of which is prioritized by SEMS.
To keep measuring probes and errors down to a mini-
mum, the measurement of all components is performed for
10 minutes, during which 10 measurements of the power
(current×voltage) are conducted per second, with a calculation
of the average time once the 10 minutes are up. Furthermore,
to ensure that the measurement of energy consumption is
reliable, the procedure is carried out five times.
All sensors and components were measured in the same
way. Table II provides the results obtained.
Table II
THE A MOU NT O F POWE R US ED BY T HE S ENS OR S AND O THE R
CO MPO NE NTS
Idle mode (W) Active Mode (W)
Arduino Mega 0.53 0.65
RPi Processor 1.25 1.86
Pixy Camera 0.12 0.74
GPS Unit 0.01 0.8
Compass kit - 0.4
Camera light 0.2 5.62
Pressure sensor - 0.01
PXFmini 0.13 0.41
XBee 0.27 0.97
A. Results of the Thruster Measurement
Thruster unit power consumption measurement was under-
taken carefully. It focuses on how and where consumption
occurred and how the measurement results could be employed
to improve performance while minimizing consumption.
As in most other AUVs, no other SEMBIO component
consumes more power than the thrusters. An ESC was used
to drive each thruster and uses power itself as well. An 11.1
V Li-Po battery was used to perform the measurement. Fig 5

shows how thrust and power consumption are correlated. The
use of SEMS is supported by how successfully the power
consumption of the thrusters was managed. Based on the
analysis of the results provided in Fig 5, it can be summarized
that the efficiency zone of the thrusters allows them to thrust
the vehicle with minimal power consumption, while in the
saturation zone the current increases without significantly
increasing the produced force. The two zones are of great
importance for SEMS to determine the optimal approach to
vehicle thrusting with a high energy efficiency and to avoid the
saturation zone. The saturation zone underwent an extension
from 73% to 100% in both directions and, therefore, its
use was not recommended. The second zone underwent an
extension from 20% to 60%, exhibiting an efficiency of over
80% in the forward direction, the same as the zone in the
backward direction, which extended from 20% to 48%. Owing
to these high efficiency zones, each thruster was capable of a
thrust of 0.25-2.3 N.
-100 -80 -60 -40 -20 0 20 40 60 80 100
PWM input signal (duty cycle %)
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Efficiency (mN/mA)
Energy Efficiency vs PWM input signal
Figure 5. Diagram showing how the energy efficiency of the thrusters
Thus, the power consumption of the thrusters in the effi-
ciency zone and in the saturation zone were respectively 2-50
W and up to 78 W.
In the efficiency zone, SEMBIO’s two forward direction
thrusters produced a combined thrust of up to 4.6 N, giving
the vehicle a speed of around 0.8 m/s (5.3 N and around 58
W produced a speed of 1 m/s). Therefore, a speed of 0.5
m/s requiring around 28 W was established for the forward
direction.
The efficiency zone was used more actively for the four
thrusters for diving. A thrust of 2 N was necessary for adequate
diving, with 0.5 N produced by each thruster necessitating
around 4.8 W, which totaled 18 W for all four thrusters in the
diving direction.
Both the thrusters and the ESC driving the thrusters ex-
hibited a ”no-load current”, meaning that each thruster used
around 1.7 W even when they were not rotating. SEMS can
address this issue by giving a stop signal to solid-state relay
(SSR) instead of ESC to deactivate the thrusters and their
drivers.
V. EXP ER IM EN TAL RES ULT S
To evaluate the performance of SEMS-Solar, and SEMS,
experiments were conducted and are illustrated in this section.
A. Experiments and Results of SEMS-Solar
To conduct the solar cell experiments, the SEMBIO cover
was used to create a SEMS-Solar prototype. To ensure its im-
permeability and for protection against environmental factors,
a transparent epoxy was used to coat the flexible solar panel.
The energy from the solar cells was extracted with the SunAir-
Plus [12] board. The solar charging control of the SunAir-Plus
board is responsible for charging the batteries, along with a
voltage booster for boosting the default 3.7-4.2 V output of
the emergency battery to 5 V. INA219 current sensors were
employed to measure the SEMS-Solar’s current and voltage.
The data from the solar logging platform was generated and
stored on an SD card.
Three different climatic conditions were measured to ensure
authentic results. The measurement was started at eight o’clock
in the morning until seven o’clock in the evening, and an SD
card was used to store the current measurement every minute
through the Arduino. When the measurement was finalized,
the processing of the data was conducted.
The correlation between the generated current or power and
direct solar irradiation is indicated by the dashed red line
in Fig 6. The accuracy and reliability of the results were
limited because the measurement was conducted at a time with
cloudy and fluctuating climatic conditions. The measurement
conducted on a sunny and warm summer day (ca. 18 ◦C) is
indicated by the dashed green line in Fig 6. By comparison
to the earlier figure, the curve that was produced provided
favorable results, yet the desired power of 1 W was not
achieved. At the ideal time, around 380 mW were generated
at 6 V.
08:00 09:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00
Time
0
20
40
60
80
100
Current in mA
0
100
200
300
400
500
600
700
Power in mW
Genera ted current and p ower of SEM S-Solar
Strong sunny day
Normal sunny day
Cloudy day
Figure 6. Solar power fluctuation in various sunny days
The measurement undertaken in ideal climatic conditions is
indicated by the blue line in Fig 6. The result fell short of the
desired 1 W, reaching only about 650 mW for a short period of
time. Therefore, owing to the climatic conditions and poorly
effective solar cells, the SEMS-solar results were suboptimal
and SEMBIO’s primary battery was not charged. Furthermore,

08:00 09:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00
Time
0
10
20
30
40
50
60
Charging time in hours
The estimated charging time of the emergency battery (400 mAh)
Figure 7. The approximate duration for charging the emergency battery
charging the emergency battery to 400 mAh was challenging
for SEMS-Solar as well.
As indicated in Fig 7, based on the premise that the
ideal results were achieved, the estimated charging time of
the emergency battery was around four hours between one
and four o’clock in the afternoon, meaning that there was a
possibility of complete charging of the battery in that period.
B. Experiments and Results of SEMS
The energy savings that can be achieved with SEMS are
discussed in the following section. A mission can be planned
better if the values of the vehicle’s components and the
amount of energy consumption are known beforehand. SEMS
was assessed by conducting a mission with SEMS and the
same mission without SEMS and subsequently comparing the
results.
1) Establishment of a Monitoring Mission: The device
user determines the manner in which a mission is developed.
For the purposes of this evaluation, a ”monitoring mission”
was undertaken to monitor and investigate a particular area
underwater, as the name implies. As can be seen in the mission
stages presented in Fig 8, which includes only one task, the
monitor task comprised four subtasks: guiding the vehicle
from the homebase to the established task coordinates based on
GPS (navigation subtask), diving the vehicle to a given depth
using pressure sensors (depth subtask), underwater exploration
based on SEMBIO’s vision (monitor subtask), and guiding the
vehicle back to homebase (return home subtask).
Depth
(Subtask 2)
Monitor
(Subtask 3)
Monitor Task
Mission
Return home
(Subtask 4)
Navigation
(Subtask 1)
Homebase
Figure 8. The stages included in the monitoring mission
Subtask creation involves the selection of components from
different categories or lists (i.e., thrusters, sensors, batteries,
processors) for use in the subtask and for generating a subtask
code with the SEMS-Mission-Planner, which is a graphi-
cal user interface (GUI) that permits mission development
over the Internet to create the subtasks, tasks, and mission.
The SEMS-Mission-Planner produces the mission packet data
(control code), which is transmitted to the robot, specifically
to the SEMS-Embedded-Board, via SEMS-Communication.
Measurement data are gathered and transmitted by the SEMS-
Embedded-Board to the SEMS-Mission-Planner in the form
of feedback so that the latter can inspect and plot them. Thus,
the horizontal thrusters, Arduino Mega, GPS, and compass
were chosen for the navigation subtasks and to guide the
vehicle back to the homebase, and the pressure sensor, Arduino
Mega, and vertical thrusters were chosen for the depth subtask.
RPi2 and its camera, camera light, and horizontal thrusters
alongside the components selected for the previous subtask
were chosen for the monitor subtask. Each subtask requires
the components to form the basis for its emergence, just
as a task requires subtasks to take form. Furthermore, task
creation and mission planning require the specification of the
start- and end-point of every subtask, in a format that must
be preprogrammed and saved in the SEMS-Embedded-Board.
For the present mission, the start- and end-points of the GPS
coordinates (longitude, latitude) of SEMBIO were established
to be (53.853363, 10.702976) and (53.854054, 10.704120),
respectively representing P0(0,0) and P1(1,1). After the diving
from the surface (start-point of the depth = D0) to a depth of
50 cm (end-point of the depth = D1), then SEMBIO began
monitoring the underwater environment for a quarter of an
hour. This subtask was connected to time, rather than to
a sensor. The creation of the monitor task comprising the
preceding four subtasks was the last stage. The mission was
created in the same manner as the task. Tasks are selected in
accordance with the objectives of the monitoring mission to
ensure that those objectives are achieved. For our experiments,
a single task was included in the monitoring mission.
Through this procedure, the whole control code was gen-
erated and transmitted to the SEMS-Embedded-Board. This
mission has only one priority case, namely, in the event of
battery capacity reduction to less than 20%, the mission has
to deactivate the vehicle, terminate every task and subtask, and
guide the vehicle back to the homebase.
2) Theoretical Evaluation of the Monitoring Mission: The
monitoring mission was evaluated theoretically for two rea-
sons: (1) to support and compare it with the real evaluation of
SEMS. The theoretical evaluations for each mission give us a
complete overview of the energy consumption of a mission; (2)
to validate that the theoretical evaluation can be an alternative
to the real evaluation, which faced challenges during the
evaluation procedures. This theoretical evaluation involved a
comparison between the overall energy consumption value of
the components employed in every subtask with and without
SEMS. For every subtask, the working premise was that the
efficiency zone of the thruster remained constant both with and

without SEMS. Therefore, the calculation of the navigation
subtask’s energy consumption (power (W) ×Time (hours)) in
the absence of SEMS was undertaken as follows.
The energy consumption of the navigation subtask was
determined by summing up the power used by the Arduino
Mega, GPS, compass, and horizontal thrusters, and the energy
consumed by the vertical thrusters, sensors, and processors in
idle mode. With SEMS, the same procedure was applied to
calculate the overall energy consumption of the navigation
subtask, which was the sum of the energy consumed by
the Arduino Mega, GPS, compass, and horizontal thrusters.
Given that SEMBIO had an energy capacity of 111 Wh,
without SEMS the navigation subtask could run for 2.87 h
(111 Wh ÷38.59 W), while with SEMS it could run for
3.72 h (111 Wh ÷29.81 W). This calculation is conducted
to determine the run time of each subtask. However, in the
presence of SEMS, a slower vehicle speed was used to ensure
that it was within the high efficiency zone of the energy
consumption, because a high speed was unnecessary as no
other subtasks were left to be completed.
Figure 9. Comparison of the power consumed by the monitoring mission
comprising one task and four subtasks in the absence and presence of SEMS
in the theoretical evaluation
3) Experimental Evaluation of the Monitoring Mission:
The procedure applied to obtain the results of SEMS for the
monitoring mission was as follows. Prior to the installation of
SEMS, the full software operating the vehicle was installed
on-board SEMBIO and the batteries were fully charged. In
the first case, each experiment was carried out in the same
manner as the experiments in the absence of SEMS. It must
be highlighted that, in the context of the actual experiment,
the term ”without or in the absence SEMS” referred to the
fact that the entire output of the SEMS-Embedded-Board was
switched on and every component was either active or in idle
mode, whereas the term ”with or in the presence of SEMS”
referred to the fact that output activation was carried out in
keeping with the control code of the ”monitoring mission”.
In the second case, SEMS was used. After receiving the
”monitoring mission” from the SEMS-Mission-Planner, the
SEMS-Embedded-Board proceeded to parse the control code
and employ SSR to switch components on or off, depending on
the requirements of the mission, tasks, and subtasks. INA219
sensors were used to measure the current and voltage of the
components, including sensors, processors, and thrusters. An
SD card stored all the measurement data on-board at a rate of
5 Hz.
Figure 10. The set-up and location of the real SEMS experiment
To ensure that the results were as reliable and valid as pos-
sible, the performance of the real experiments in the Waknitz
river were repeated several times. No complex equipment was
needed to conduct the evaluation of SEMS, the vehicle with
a laptop and fishing reel (to pull the vehicle in case of any
faults) being sufficient (Fig 10).
The procedure adopted in the theoretical evaluation was
employed in the real experiment as well in order to avoid any
complication or problems. In the theoretical evaluation, two
scenarios were enacted, namely, one mission with a single task
and four subtasks being performed with and without SEMS,
respectively. However, the experimental evaluation of SEMS
encountered a number of issues. One issue was that the first
subtask (i.e., the navigation subtask), was not successfully
performed by SEMBIO, failing to reach the intended location
at the surface of the water by relying on GPS. To put it
differently, the vehicle did not follow the ideal path between
P0 (53.853363, 10.702976) and P1 (53.854054, 10.704120).
The solution to this issue was to ignore the fault, permitting
SEMBIO to move freely at the surface of the water. However,
the subtask was connected to a specific time, after which the
vehicle was submerged to an established depth for a set amount
of time as well.
SEMS measured a number of different parameters, includ-
ing the total power (W), current (A), battery voltage (V),
battery capacity (Wh), and the diving depth (cm). The data
formed the basis for the calculation of the energy consumed.
The average total power used by the navigation subtask (44.5
W), which means that for one hour the energy consumption
44.5 Wh. There was a difference of about 6 W between the
theoretical evaluation and the real evaluation (38.59 W vs
44.5 W). Potential reasons for this difference include the fact
that running some of the components in idle mode was not
successful or the fact that the battery voltage was changed
from 11.1 V in the theoretical evaluation to 11.65-12.2 V in the
real evaluation. For the following three subtasks, an identical
approach as in the case of the navigation subtask was adopted.
The depth subtask consumed about 26.20 Wh. Compared to
the other subtasks, the monitor subtask was less complicated. It
involved moving the vehicle horizontally at the desired depth

of 0.5 m. During that time, the energy consumption of the
vehicle was around 33.26 Wh, more details in [13].
Similar to the first subtask of navigation, the final subtask
of return to the homebase involved floating the vehicle to
the surface and moving it from its location to the homebase.
This subtask was performed in a different way, depending on
whether SEMS was used or not. In the absence of SEMS, the
performance of the subtask did not differ from the performance
of the navigation subtask.
In the final part of the experiments, SEMS was used, and its
evaluation involved the performance of the same procedures,
with the exception that the monitoring mission was initiated
after the battery was completely charged and SEMS was
activated.
The real experiments with and without SEMS
A Navigation Depth control Monitor Return home
Subtasks
0
10
20
30
40
50
The totol power in W
without SEMS
using SEMS
Figure 11. Comparison of the results of the real experiments with and without
SEMS in terms of the energy consumption for a monitoring mission with one
task and four subtasks
Table III
OVE RVIE W OF TH E RE SULT S OF TH E RE AL EX PE RIM ENT S WI TH AN D
WITHOUT SEMS, TH E TIM E OF E NER GY CO NS UMP TI ON IN A LL S UBTA SKS
IS CALCULATED ACCORDING TO THE ENERGY CAPACITY OF SEMBIO,
WHICH IS 111 WH
Mission without SEMS Mission with SEMS
Subtasks Power Time Power Time Percentage gain
using SEMS
Navigation 44.50 W 2.49 h 32.45 W 3.42 h 37.35 %
Depth control 26.20 W 4.23 h 20.42 W 5.43 h 26.87 %
Monitor 33.26 W 3.34 h 31.70 W 3.50 h 4.79 %
Return home 42.60 W 2.61 h 17.82 W 6.23 h 138.7 %
Total sum 146.5 W 0.76 h 102.4 W 1.08 h 42.11 %
Fig 11 presents the differences identified when the real
experiments with and without SEMS were compared. Ta-
ble III provides an overview of all the results. Identical to
the results obtained in the theoretical evaluation, the use of
SEMS reduced the energy consumption in all subtasks in the
real experiments, and the results of the real evaluation were
close to the theoretical evaluation, with only slight differences.
The use of SEMS was highly advantageous, not only helping
to secure a gain of more than 40 % in energy usage, but also
to better manage the different vehicle components.
VI. CONCLUSION AND FUTURE WO RK
The present paper focused on the discussion of the SEMS
system as a viable option for managing SEMBIO’s energy con-
sumption, reducing the consumption, and therefore prolonging
the mission time by controlling the restricted on-board energy
source. SEMS can achieve these objectives through its four
interconnected and inter-collaborative parts. The performance
and analysis of SEMS and SEMS-Solar experiments and their
results were also undertaken in this paper. It is evident from the
findings that, when utilizing SEMS, the micro AUV consumes
less power up to 40%. This resulted in extended battery life
and the vehicle accomplishing more tasks before it had to be
returned to the base for a battery recharge.
Future research should focus on expanding the SEMS con-
cept to integrate a range of AUV types and produce a manifest
file mapping the AUV components with main board pins to
verify the components prior to mission launch. The subject
of smart energy management also demands consideration
so that vehicle components can be controlled according to
environmental conditions and mission characteristics.
REFERENCES
[1] J. Yuh and H.-T. Choi, “Unmanned underwater vehicles,” Wiley Ency-
clopedia of Electrical and Electronics Engineering, 2015.
[2] D. Crimmins, C. Patty, M. Beliard, J. Baker, J. Jalbert, R. Komerska,
S. Chappell, and D. Blidberg, “Long-endurance test results of the solar-
powered AUV system,” in OCEANS 2006, pp. 1–5, 2006.
[3] T. Hyakudome, “Design of autonomous underwater vehicle,” Interna-
tional Journal of Advanced Robotic Systems, vol. 8, no. 1, p. 9, 2011.
[4] E. Sahin, S. Girgin, L. Bayindir, and A. E. Turgut, “Swarm robotics.,”
Swarm intelligence, vol. 1, pp. 87–100, 2008.
[5] M. A. Joordens, T. Shaneyfelt, K. Nagothu, A. Jaimes, and M. Jamshidi,
“Applications and prototype for system of systems swarm robotics,” in
Systems, Man and Cybernetics, 2008. SMC 2008. IEEE International
Conference on, pp. 2049–2055, Oct 2008.
[6] E. S¸ahin, “Swarm robotics: From sources of inspiration to domains of
application,” in International workshop on swarm robotics, pp. 10–20,
Springer Berlin Heidelberg, 2004.
[7] U. Behrje, C. Isokeit, B. Meyer, and E. Maehle, “A robust acoustic-
based communication principle for the navigation of an underwater robot
swarm,” in 2018 OCEANS-MTS/IEEE Kobe Techno-Oceans (OTO),
pp. 1–5, 2018.
[8] A. Amory and E. Maehle, “Modelling and CFD simulation of a micro
autonomous underwater vehicle SEMBIO,” in OCEANS 2018 MTS/IEEE
Charleston, pp. 1–6, Oct 2018.
[9] A. Amory and E. Maehle, “SEMBIO - a small energy-efficient swarm
AUV,” in OCEANS 2016 MTS/IEEE Monterey, pp. 1–7, 2016.
[10] DIGI, “XBEE 2.4 GHz series of modules.”
http://www.digi.com/products/xbee-rf-solutions/2-4-ghz-modules,
2015. [Online; accessed 27-Sep-2015].
[11] Erlerobotics, “PXFmini Board.” http://docs.erlerobotics.com/brains/pxfmini,
2015. [Online; accessed 12-Sep-2016].
[12] Switchdoc, “SunAirPlus solar power controller.”
http://www.switchdoc.com/sunairplus-solar-power-controllerdata-
collector/, 2015. [Online; accessed 17-Apr-2019].
[13] A. Amory, “SEMBIO - an energy efficient micro underwater robot for
swarm applications,” Dissertation, University of Luebeck, 2018.