On-demand Electric Thrust Assistance for Muscle-powered Watercraft:

PADDELEC

Ren

e Budich

, Stephan Zipser, Jakob Doblaski and Jonas Seidel

Institute for Electric Mobility, University of Applied Sciences Dresden, Friedrich-List-Platz 1, Dresden, Germany

Keywords:

Paddelec, Demand-oriented Support, Electric Mobility, BEV, Canoe, SUP, Watercraft, Boat Simulation.

Abstract:

In the project PADDELEC, the implementation of an intelligent paddle (iPaddle) for controlling an electric

assisted canoe was realized. This paddle is able to record the paddle force F

of the blade with the help of

measurement technology integrated in the paddle shaft. The recording of the paddle stroke is used as a control

variable for operating an electric auxiliary drive in the boat. This is intended to provide on-demand thrust

assistance to the athlete, analogous to an electriﬁed bicycle (pedelec). Existing drive solutions on motorized

boats offer the athlete the option of ﬁrmly specifying the desired thrust via a thrust lever or tiller. The paddelec

has the aim to provide intelligent and dynamic thrust assistance to the athlete as needed. This is intended to

preserve a natural driving feel, despite the auxiliary drive. A simulation model of the longitudinal dynamics

of canoes, was developed to investigate various assistance strategies. The models were validated by extensive

real driving tests. For this purpose, special hardware and software tools had to be created, which will be further

developed in the future. With the simulation and the evaluation of the practical testing, it could be shown that

thrust support for canoes can be very useful. The correct support strategy and support performance can vary

greatly depending on which water conditions the boat will use and the sporting activity, so there is still a need

for more research in this area.

1 INTRODUCTION

Water sports become more and more popular around

the world. For example, there has been a steady and

substantial increase in water touring by canoe (e.g. in

countries like Canada, Sweden and Germany). The

use of SUPs (Stand-Up-Boards) is also increasing for

some years now (BMWi, 2013).

In many cases water sports are performed at calm,

near-natural coastal, inland waters or low-current

rivers. However, it remains an outdoor sport where

wind and water current (and/or tide) often require a

powerful athlete.

To make this water sport available to a larger user

group, a battery-electric propulsion support is an ob-

vious suggestion. In case of bicycles, the idea of

battery-electric support is known as pedelec (pedal

electric cycle) and has been extraordinarily commer-

cially successful in Europe.

As known, real live disturbances like wind and

water current may affect the results of practical ﬁeld

tests substantially. For this reason, physically oriented

models were developed to describe the response of a

https://orcid.org/0000-0002-9227-6834

boat to human and motor propulsion forces analyti-

cally. Model simulations in comparison to ﬁeld tests

are also presented later.

The project was largely driven by engineering stu-

dents, who developed and implemented technical so-

lutions for measuring and controlling the boat-drives

and tested them in driving trials.

2 THE IDEA OF PADDELEC

The objective of the paddelec project is to develop a

system which provides on-demand electric thrust as-

sistance analogous to the established electric pedal as-

sistance for bicycles: the propulsion input of the user

is measured and supported electrically.

The basic principle of both technologies is identi-

cal.

In both cases, the mechanical input of the user

(or less technical the desire for assistance) is mea-

sured with a torque or force sensor. In the case of ca-

noeing, this is done with a paddle equipped with sen-

sors. The measurements are processed by a control

This is why the name analogy to the pedelec was de-

liberately chosen.

Budich, R., Zipser, S., Doblaski, J. and Seidel, J.

On-demand Electric Thrust Assistance for Muscle-powered Watercraft: PADDELEC.

DOI: 10.5220/0011380600003321

In Proceedings of the 10th International Conference on Sport Sciences Research and Technology Support (icSPORTS 2022), pages 17-28

ISBN: 978-989-758-610-1; ISSN: 2184-3201

Figure 1: Block diagram of a paddelec.

unit, which drives the propulsion motor controller.

Depending on the amount of power applied, the ve-

hicle can thus be controlled by the driver.

The main focus is to provide a driving experience

that is as natural as possible without the use of addi-

tional peripherals such as a throttle.

In a canoe, the propulsion force or power is trans-

mitted with the help of a paddle. When paddling, the

paddle is pushed against the water to move the canoe

forward. To do this, the paddle blade is inserted into

the water and the paddle shaft is pushed against it.

Fig. 1 shows the basic structure with the aid of

a block diagram. The blocks shown in red are rep-

resenting additional components which are necessary

for implementing demand-oriented thrust support.

The paddle (Fig. 1 - ➁) exerts a force F on the hull

of the boat (Fig. 1 - ➀) so that it moves longitudinally

in the water. It should be noted here that the mea-

sured paddle force F

can never be used to 100 % for

propulsion (paddle force F

̸= propulsive force F

)

(see also section 5.2.2).

In order to realize demand-oriented support, the

force component in the direction of travel must be

recorded. This is done with a special measuring pad-

dle (iPaddle) which, among other things, has inte-

grated sensors and strain gauges for position and force

diagnostics. The recorded sensor values can be used

to control an electric motor dynamically. The battery,

motor controller, motor and propeller respectively jet

are parts of the power train unit, which can be de-

signed in different ways (Fig. 1 - ➂).

3 STRUCTURE OF THE TEST

VEHICLE

3.1 Boat Hull Grampus I

The Grampus I (a two-man kayak) from the Kaitts

company shown in Fig. 2 was used for the test setup.

The hull is made entirely of plastic, which makes

the boat very robust. This also makes it easy to imple-

ment modiﬁcations. The kayak has a wide hull cross-

section, which provides high stability against tipping

on the water and increased safety during test runs. In-

side there is enough space for a traction battery, con-

trol unit and measurement equipment. The rear part

of the kayak is sealed off watertight, allowing water-

sensitive components and measurement technology to

be accommodated via a hatch.

Figure 2: Boat hull - Grampus I (Kaitts Ltd., 2022).

icSPORTS 2022 - 10th International Conference on Sport Sciences Research and Technology Support

3.2 iPaddle - Measuring Paddle

Fig. 3 shows a part of a classic double paddle typi-

cally used in popular sports. The developed iPaddle is

equipped with force and position sensors and a wire-

less data transmission to the control unit.

The tensile force applied by the paddler causes

a deformation of the paddle shaft. This bending is

detected by a strain gauge and transmitted to the

ADC (analog digital converter) as an electrical volt-

age change. The ADC ampliﬁes the signal and con-

verts it into a digital signal, which is transmitted to

the control unit in the boat (Dr. Budich and Dabrazzi,

2020).

Figure 3: FEM analysis of the measuring paddle (H

ase,

2020).

For the application of the strain gauges on the

paddle, the area of highest stress should be selected.

Therefore, the double paddle was statically loaded

and then examined. The resulting stress curve over

the paddle shaft was calculated before using a FE cal-

culation (Finite elements method). It can be seen that

the maximum bending moments and thus the maxi-

mum tensile and compressive stresses lie in the area

of the grip point closest to the paddle blade (Fig. 3).

In order to determine all forces correctly, an exact

position measurement is required, which is currently

still under development.

For an initial estimate, it is sufﬁcient to use a pad-

dle factor to determine the propulsive force F

T hrust

) from the measured force F

Blade

In further development we can use the measured

paddle angle α to do the correction more accurately

(see equation 1).

(α) = F

· K(α) (1)

3.3 Power Train

The basic structure of the paddelec is shown in Fig. 4.

The structure consists of the components: Boat body

(canoe) ➀, iPaddle ➁, battery ➂, control unit ➃ and

the drive unit ➄ (e.g. water jet drive).

Figure 4: Basic structure paddelec (P

oschmann, 2020).

The electric auxiliary drive generates the desired

additional thrust assistance through appropriate con-

trol. A waterjet propulsion system is favored as the

drive unit because, unlike a propeller, it can be easily

and safety integrated into the hull of the boat. It is also

assumed to have a lower impact on the environment.

The waterjet drive for the test vehicle is made up

of several individual components. All components

were designed for additive manufacturing and 3D

printed from polylactide plastic. The waterjet propul-

sion system is screwed to the fuselage and sealed us-

ing sealing compound. Fig. 5 shows the designed jet

drive in installation position, Fig. 6 the underside of

the hull with outlet opening and protective ﬁns.

Figure 5: Constructed jet drive in installation position with

motor and belt drive (Hauptmann, 2020).

Figure 6: Bottom of the fuselage with protective ﬁns and

deﬂection of the outlet (Hauptmann, 2020).

In addition, investigations were carried out with

On-demand Electric Thrust Assistance for Muscle-powered Watercraft: PADDELEC

Figure 7: KDN-W703000 (Kedean, 2022).

a commercial drive solution (KDN-W703000) from

Kedan (Fig. 7). A brushless DC motor is used here

as the electric machine, while an impeller is used as

the drive. These two components are installed in an

aluminum housing as a combination. By using an im-

peller, the risk of injury is minimized in contrast to a

propeller. Table 1 lists the available technical data for

the drive.

Table 1: Technical data Kedean motor (Kedean, 2022).

technical characteristics values

max. operating voltage 50 V

max. continuous current 75 A

max. power 4000 W

max. thrust 240 N

length 401 mm

weight 2.5 kg

impeller diameter 60 mm

3.4 Traction Battery

A battery solution from Aentron Energy Solutions

was chosen to supply the system with energy. The

company specializes in energy storage systems for the

maritime sector. Important characteristics are shown

in table 2.

The battery has an integrated battery management

system (BMS), which provides overcharge and deep

discharge protection, as well as temperature moni-

toring. Its robust aluminum housing and protection

against dust and water make it a good solution for car-

rying out the measurement runs.

Table 2: Technical data battery (Aentron, 2022).

technical characteristics values

capacity 20.3 Ah

energy 1023 Wh

continuous discharge power 3000 W (@ 60 A)

nominal voltage 50.4 V

operating voltage 42 V- 56 V

weight 9 kg

protection class IP66

3.5 Electronic Control Unit - ECU

The developed control unit is based on an Arduino

environment. The ECU acquires measurement data,

processes it and sends corresponding signals to a mo-

tor controller. The necessary software algorithms

were implemented with Matlab Simulink. Its also

possible to log sensor data. The mode of operation

is shown schematically in Fig. 8.

Figure 8: Basic scheme of ECU (Doblaski, 2021).

3.6 Housing and Peripherals

The control unit and the necessary peripherals must

be securely housed on the boat. For this purpose, a

housing was developed which protects the technology

used from penetrating dirt and water (Fig. 9).

The D-Sub 9 connector ➀ serves as input for the

paddle’s measurement data. The XT60 connector ➁ is

used to connect the battery. The plastic cable gland ➂

is used to guide the motor cable to the motor con-

troller. The four-pole connector ➃ is used to record

the measured data from the battery. A switch ➅ for

switching the operating voltage of the Arduino on and

off was mounted on the top. Two LEDs ➆ serve as

status and error indicators. Furthermore, a rotary in-

cremental encoder ➇ and a potentiometer ➈ were at-

tached to the top. With these, it is possible to change

parameters while driving. In this way, different sce-

narios can be tested quickly during the measurement

runs. To make the parameter changes visible, a dis-

play ➉ was also mounted in the housing.

Figure 9: Housing for the electronic components (Doblaski,

2021).

icSPORTS 2022 - 10th International Conference on Sport Sciences Research and Technology Support

4 SUPPORT STRATEGIES FOR

THE PADDELEC

4.1 Conceptual Approaches

There are various strategies for demand-oriented sup-

port. Depending on the driver’s wishes and objec-

tives, the drive system implements a thrust depend-

ing on the support strategy selected. In the follow-

ing, three selected basic strategies will be examined

in more detail.

Figure 10: Basic forces for analysis support strategies.

The propulsive forces acting on the boat are shown

in Fig. 10. They are divided into paddle forces F

and motor forces F

(orange arrows). The sum of the

two forces is the total thrust force F

(green arrow).

This force acts against the resistance force F

(black

arrow).

4.2 Strategy 1 - Constant Support

In this strategy, the user is assisted by the drive with

constant force. For this purpose, the force applied by

the user to the paddle is averaged over several pad-

dle strokes. The calculated force is then applied con-

stantly by the drive for support (mode: tailwind).

Fig. 11 shows the individual force curves over

time. The red line shows the progression of the force

applied to the paddle.

Figure 11: Strategy 1 - Constant support scheme.

The green line is the additional force F

applied by

the motor

. The resulting force F

from the paddle

and motor is indicated by the blue line.

The aim of this strategy is to be as universal as

possible for all requirements. It offers a compromise

between dynamics and efﬁciency.

4.3 Strategy 2 - Direct Support

This strategy is based on the paddling rhythm of the

user. Through the direct use of assistance, each pad-

dle stroke is perceived as more powerful and the reac-

tion of the boat as particularly dynamic

The strength of the assistance is aligned to the

force measured at the paddle (F

). The propulsion

system replicates the force curve over a paddle stroke.

As a result, the pushing force F

should act syn-

chronously with the paddle force F

. The Fig. 12

shows the corresponding curve progression. It is spe-

cial that the motor force has the same qualitative pro-

ﬁle of the paddle force.

This strategy corresponds to the approach of a

pedelec (elec. bicycle). Due to the simultaneous

thrust of the drive, the paddle stroke should be po-

tentiated and thus be perceived as more powerful.

Figure 12: Strategy 2 - Direct support scheme.

4.4 Strategy 3 - Valley Support

In this strategy, the goal is to achieve high efﬁciency

by moving the boat as smoothly as possible. The

thrust of the motor F

is used to supplement the pad-

dling force F

so that there is always a constant force

on the boat. The maximum force to be supplemented

is determined by the driver. Fig. 14 shows the pro-

gressions of the forces during paddling.

Calculated from the averaged force of the paddle.

The athlete has the impression of being much stronger.

On-demand Electric Thrust Assistance for Muscle-powered Watercraft: PADDELEC

Figure 13: Basic structure of the simulation.

The advantage of this strategy is that the boat

causes less dynamic motion and can therefore be

driven more efﬁciently. The motor takes over the

power requirements to propel the boat smoothly. The

goal of this strategy is to eliminate speed ﬂuctuations

(∆v) as much as possible.

Figure 14: Strategy 3 - Valley Support scheme.

5 INVESTIGATING SUPPORT

STRATEGIES USING A

SIMULATION APPROACH

5.1 Aims and Principle Approach of the

Simulation

One way to test the strategies mentioned above is to

conduct real-life tests with the canoe. However, this

requires a lot of technical and logistical effort. An-

other approach is to investigate the various support

strategies using a holistic simulation approach. This

makes it possible, for example, to identify errors in

the actuation strategies in advance. This approach is

signiﬁcantly more efﬁcient and saves resources.

Accordingly, the goal is to develop a physically

correct simulation model, for the investigation of the

support strategies, in order to verify the presented ap-

proaches. Furthermore, the developed strategies can

be transferred by a rapid prototyping approach, into

C-code, which can be used directly in the microcon-

troller.

The simulation models are ﬁrst developed with the

assumption of ideal conditions. This means that there

are no disturbing inﬂuences in the model. In reality,

however, these always exist and include, for exam-

ple, the inﬂuence of wind, currents and waves. Some

of these inﬂuences can be approximated by constants,

but deviations from the real system inevitably remain.

These have to be estimated and evaluated.

The simulation and modeling of the entire sys-

tem is realized with the program Matlab/Simulink

from Math Works. The simulation consists of differ-

ent blocks, each representing a sub-component of the

overall system. The basic structure of the simulation

model shown in Fig. 13 consists of several subsystems

and is based on real conditions.

For example, the ”Human” simulation block

serves as input for the iPaddel model. The output vari-

able paddle force is passed on as inputs to the ”ECU”

and ”Boat hull” blocks. The ECU has the task of

generating a PWM signal based of the revived sen-

sor data, which is used to control the motor in Block

”Power train”.

The simulation results, especially the speed of the

boat, can be validated with real measured values (e.g.

with a GPS-sensor).

5.2 Realization of the Simulation

5.2.1 Model of the Boat Hull

The boat is a rigid body. To set it in motion a force

must be applied:

a =

(2)

where a is the acceleration, F is the total force acting

on the body, and m is the mass of the body including

the mass of the persons.

icSPORTS 2022 - 10th International Conference on Sport Sciences Research and Technology Support

Figure 15: Signal ﬂow diagram in Simulink (Seidel, 2022).

The speed of the boat is then obtained by integrat-

ing the acceleration to give:

v =

adt + v

(3)

The force F

Forward

) results from all the partial

forces (paddle force F

, motor force F

, resistance

force F

and thrust force F

) acting on the boat to be:

= F

+ F

− F

= F

− F

= m · a (4)

In a canoe, the force F

results from the movement

of the paddle in the water. An additional force F

generated by a motor used. These two forces combine

to form the thrust force F

. This is countered by a

resistance force F

, which dependents, among other

things, on the shape of the boat. Fig. 10 shows this

situation on a canoe.

In order to create a simulation, it is therefore nec-

essary to ﬁnd out which input forces act on the boat

through the paddle and motor and how the boat be-

haves in terms of the resistances that occur.

Modeling the physical behavior of a watercraft

is very complex. In order to model the drag force

as realistically as possible, a large number of phys-

ical equations and parameters must be considered.

The total resistance of the boat is made up of several

partial resistances, some of which are very complex

to determine (e.g. skin friction resistance and pres-

sure resistance, which can be further divided in wake-

making resistance, naked hull skin friction resistance

appendage skin friction resistance and so on).

The model can quickly become inaccurate and un-

realistic because there are a large number of submod-

els for which assumptions have to be made. There-

fore, the modeling is additionally supported by metro-

logical investigations, e.g. by measuring the deceler-

ation behavior. The model was created in Simulink

according to equations 2, 3 and 4 with the signal ﬂow

diagram shown in Fig. 15.

For the determination of the boat resistance, the

equation for the drag coefﬁcient was used:

cw =

2 · F

ρ · A · v

(5)

The cw value, also called the drag coefﬁcient, is

composed of the density of the medium ρ, the face

area A, the velocity v of the incident ﬂow of the

medium and the drag force F

In this context, the frontal area is the area that the

respective medium, in this case water or air, impinges

on in the direction of travel. The density of water is

assumed to be constant at 1000 kg/m

and the density

of air is assumed to be constant at 1.2 kg/m

Two ﬂuids act on the paddler and the boat, on the

one hand the air and on the other the water. Thus, after

rearranging equation 5 and considering both ﬂuids at

once, the total drag force F

R,total

R,t

) results in:

R,t

= F

R,w

+ F

R,a

(6)

Where the subscripts a and w stand for air and water

and t for total, respectively. This gives the equation:

R,t

· ρ

· cw

· A

· v

· ρ

· cw

· A

· v

(7)

It is assumed for modeling purposes that there are

no perturbations such as current or wind acting on the

boat. The velocities of the inﬂow from air and from

water are therefore equal to the velocity of the boat.

Equation 7 can thus be simpliﬁed:

R,t

· v

· (ρ

· cw

· A

+ ρ

· cw

· A

) (8)

It is difﬁcult to capture the frontal area of the boat

including the athlete and paddle

. For this reason, the

factors K

and K

were used, which combine the un-

known drag coefﬁcient and the unknown frontal area

of the respective medium. This results in the equation

of the total resistance to be:

R,t

· v

· (ρ

· K

+ ρ

· K

) (9)

After factoring out ρ

, we get:

R,t

· v

· ρ

· K (10)

transformed, where:

K = K

· K

(11)

The parameterization of the model was deter-

mined using the deceleration behavior of the boat,

similar to the usual procedure in automotive engineer-

ing. The magnitudes of the driving resistances are di-

rectly dependent on the velocity. In order to deter-

mine this relationship over the entire speed curve, a

deceleration test was carried out.

The area changes continuously during the ride, e.g.,

due to the continuous movement of the paddle, the sitting

posture of the paddler, or the slight lifting out of the boat at

higher speeds.

On-demand Electric Thrust Assistance for Muscle-powered Watercraft: PADDELEC

Figure 16: Acceleration and deceleration curve (Seidel, 2022).

The boat was accelerated to a given speed (e.g.

8 km/h) and then switched off. The drag forces acting

on the boat result in a reduction of the boat speed until

it comes to a standstill. Provided that no disturbance

variables such as wind and current inﬂuence the be-

havior of the boat, the force acting on the boat can be

calculated from the determined acceleration (respec-

tive negative acceleration).

The mass of the test vehicle was determined be-

forehand. Thus, the acting force F can be calculated

at any time. The force and velocity can then be re-

lated via the time stamps. This experiment is carried

out under the assumption that the resistance behaves

identically at the same instantaneous speed - regard-

less of whether the boat is accelerating or moving at

a constant speed. To obtain reliable values, the exper-

iment is repeated several times and the mean value is

calculated. With this procedure, the resistance from

the speed could be determined.

Following the driving resistance determination of

motor vehicles, the velocity curves of several deceler-

ation curves of the boat were recorded for the deter-

mination of the force F

R,t

. The acceleration was then

calculated from the recorded velocities by derivation.

The acceleration and roll-out curves can be ﬁtted

2nd or 3rd order polynomials. Therefore, an approxi-

mation of the real data is possible. An example com-

parison of the ﬁtted and acquired data can be seen in

Fig. 16. By the conversion into a continuous function

a derivation is now possible. (Doblaski, 2021; Seidel,

2022; Vikulin, 2021)

5.2.2 Model of the Paddle

Once the boat model has been validated, a ”standard

paddle stroke” for amateur athletes must be deﬁned

and converted into a functional block for comparative

studies. To do this, the maximum force must ﬁrst be

determined and speciﬁed.

Fig. 18 shows forces on kayak paddles during use

by competitive athletes. The forces are plotted in

newtons for women’s (F) and men’s (M) rides at dif-

ferent paddle cadences in strokes per minute (spm).

The force of a female athlete at 60 spm, corresponds

to forces of up to 126 N. This deﬁnes the maximum

values.

The acquired force on the paddle is not fully used

for propulsion. The captured force must be corrected

with a paddle function. For initial assumptions, a pad-

dle factor should sufﬁce for now.

The paddle factor depends, among other things,

on the blade shape. Symmetrical paddles have an ef-

ﬁciency factor of up to 74 %, asymmetrical paddles

up to 89 %. The paddle used for the tests is highly

asymmetrical. Both the paddle angle and the paddle

distance from the boat have also a large inﬂuence on

the effective propulsive force (Jackson, 1995; Vikulin,

2021). To reduce the number of unknown variables, a

very accurate paddle guidance is assumed.

To capture the correct paddle factor, real measure-

ments of the paddle are given to the validated boat

model and ﬁtted until the curves match sufﬁciently

accurately. Fig. 17 shows the measured data of the

left and right paddle as well as the resulting force

acting on the boat through the paddle as propulsive

force. The measurements were made on a lake near

by Dresden under good weather conditions (only light

icSPORTS 2022 - 10th International Conference on Sport Sciences Research and Technology Support

Figure 17: Boat hull validation.

wind). Special measurement equipment from Lord

ZSE Electronic was used for the tests. The speed

recording is based on a satellite-based measurement

like GPS.

Figure 18: Force applied to kayak paddles by competi-

tive athletes by paddle frequency and gender (Gomes et al.,

2015).

5.2.3 Model of the Power Train

The control of the motor is realized by pulse width

modulation (PWM). The calculated duty cycle of the

desired PWM signal is transferred to the motor con-

troller via a microcontroller. This is housed with

the battery and the drive in the drive train subsys-

tem (Fig. 1). The motor controller controls the motor,

which generates a force in the longitudinal direction

via the impeller on the drive shaft. As a further input,

the boat subsystem also has the force of the paddle,

which also drives the boat.

From the calculation for the jet model, it is known

that the thrust depends on the rotational speed as well

as the difference between the ejection and inﬂow ve-

locities. In order to determine this relationship in a

way that could be used for the model, several test runs

were carried out and the results were then stored in the

form of a characteristic diagram. This means that the

drive model can be treated as a black box.

5.3 Validation of the Simulation

The vehicle model consists of subsystems of dif-

ferent types and behavior, which requires a sepa-

rate approach (simulation approach) depending on

the part model (sub-model). To meet the require-

ments of a complete vehicle simulation (integration

of ECUs, consideration of physical behavior, use of

analog controllers), a hybrid simulation approach or

co-simulation must be used.

Fig. 20 shows the complete simulation model,

based on the block diagram in Fig. 1. In addition, the

”ECU” and ”User Input” block has been added. The

control algorithms are implemented in the ”ECU” and

possible user inputs are processed in the ”User Input”

block.

To validate the simulation, the force of the paddle

acting for propulsion is applied to the boat model and

then the resulting simulated speed is compared with

the real recorded speed. Since the recorded force of

the paddle (see also section 5.2.2) does not act directly

as a propulsive force, an adjustment was made using

a paddle factor.

The aim is to simulate the real speed as accurate

as possible using the Simulink model. If the course

of the real speed matches the simulated speed sufﬁ-

On-demand Electric Thrust Assistance for Muscle-powered Watercraft: PADDELEC

0 100 200 300 400 500 600

time (s)

100

0 100 200 300 400 500 600

time (s)

0.5

1.5

Figure 19: Htw Dresden paddle cycle for analyzing paddelec support strategies.

ciently accurate, the simulation model has been pa-

rameterized correctly and can be used for further in-

vestigations.

Fig. 17 shows an example of a measurement run

that was used to validate the simulation model. The

respective left and right paddle forces are shown at

the top left and right. These are the measured force

on the paddle. In the lower picture, the measured

velocity is shown in green (V

meas

) and the result of the

simulated velocity in blue (V

sim

In the initial range of the measurement (1-60 sec-

onds) clear deviations can be seen, which can be ex-

plained by wind inﬂuence. The remaining course of

the simulated speed agrees very well with the mea-

sured one. Thus, the model can be accepted as vali-

dated for this individual case.

Figure 20: Structure of the overall simulation.

5.4 Simulation Results

5.4.1 Paddle Cycle for Testing Support

Strategies

A paddle cycle of about 10 minutes is used to in-

vestigate the different strategies. To vary the pad-

dle strokes, the maximum value is scaled. The pad-

dle frequency is not changing over time. Fig. 19

shows the progression. First, the athlete paddles with

the synthetically generated simulated ”norm paddle

stroke for amateur athletes” for about 120 seconds

with 100 % of the speciﬁed maximum force of 126 N

). Afterwards there is a ca. 2-minute break. Then

the athlete paddles for 70 seconds with 60 %, makes a

break of 40 seconds and returns to paddle with 25 %

of the speciﬁed maximum force. After another short

break, there are 3 fast paddle cycles of ca. 10 seconds

followed by a constant phase of 50 % (90 seconds).

The cycle runs until the canoe stops.

The lower section of the ﬁgure shows the simu-

lated behavior of the boat without the inﬂuence of

disturbance variables or support forces. A maximum

speed of approx 2.1 m/s is achieved.

5.4.2 Comparison of the Strategies

Fig. 21 (upper plot) shows the velocity proﬁles with

the application of all three strategies explained before.

The paddle proﬁle from Fig. 19 was used for the sim-

ulation. The lower plot shows the additional forces

added by the motor. The green curve (dotted line)

shows the velocity proﬁle without additional support.

It is noticeable that the thrust assistance has a

icSPORTS 2022 - 10th International Conference on Sport Sciences Research and Technology Support

0 100 200 300 400 500 600 700 800

time (s)

0.5

1.5

0 100 200 300 400 500 600 700 800

time (s)

100

150

Figure 21: Simulated speed curves of different support strategies.

50 51 52 53 54 55 56

time (s)

1.9

2.1

2.2

2.3

51 52 53 54 55 56

time (s)

100

150

Figure 22: Zoomed curves from ﬁgure 21.

greater effect at lower speeds. A signiﬁcant speed

difference at maximum paddle force is hardly notice-

able. In order to see a large change here, a more

powerful power train is needed, since the speed en-

ters quadratically into the resistance calculation (see

also equation 10). However, this does not make sense

for the envisaged canoeing, since it also means that a

larger battery is needed and correspondingly heavier

drive components. An acceptable compromise must

be found here.

In Fig. 22, the time course from graph 21 is

zoomed in. So that the different behavior of the strate-

gies can be seen better. The purple curve v

sim,V S

shows a very smooth speed curve, which is achieved

by a correspondingly dynamic control of the drive

in the valleys (F

M,V S

). The red curve v

sim,CS

corre-

sponds to the original curve without support but with a

given offset. The blue curve v

sim,DS

shows the largest

changes in velocity and is therefore also the worst in

terms of energy. However, it can be assumed that this

mode is perceived as very dynamic and met with great

approval.

6 CONCLUSIONS

The fundamentals of an intelligent paddle (iPaddle)

and its suitability as a control device for electric aux-

On-demand Electric Thrust Assistance for Muscle-powered Watercraft: PADDELEC

iliary drives were investigated.

The behavior expected in advance in terms of

thrust could be conﬁrmed in principle. In the future,

the use of an optimized waterjet drive is planned. The

main aim is to reduce propulsion losses and to inte-

grate the propulsion system into the hull of the boat

in a form-ﬁt manner.

Extensive hardware and software tools were cre-

ated for the metrological investigation of an electri-

cally assisted canoe.

A new simulation model of the longitudinal dy-

namics was developed. The drive train and driving

resistance subcomponents were modeled in the form

of a black box using the measurement data from driv-

ing tests. This can be used among others to test the

support strategies in the laboratory.

The realistic behavior of the overall simulation

was examined and evaluated using an extensive test

scenario. The basic strategies can now be further sup-

plemented and examined with the aid of the simu-

lation. The theoretically elaborated strategies could

partly be tested practically on the water.

By means of recorded data of the measurement

runs, further investigations can be carried out on the

boat, e.g. the inﬂuence of disturbance variables.

With the evaluation of the practical testing, it

could be shown that some support strategies behave

signiﬁcantly different in reality than was assumed the-

oretically.

The use of a position sensor for the iPaddle is cur-

rently being investigated to measure the position in

space more accurately. This will make it possible to

develop more precise strategies and implement new

safety features, such as preventing the motor from be-

ing supported in certain situations like special course

corrections.

REFERENCES

Aentron (2022). Energy solutions.

BMWi (2013). Water tourism in germany.

Doblaski, J. (11/2021). Investigation of model-based ac-

tuation strategies for an electrically powered kayak.

Diploma thesis, HTW Dresden, Dresden.

Dr. Budich, R. and Dabrazzi, E. (2020). Drive system for

a watercraft that can be moved by muscle power, and

method for controlling a drive system.

Gomes, B. B., Ramos, N. V., Conceic¸

ao F, A. V., Sanders,

R. H., Vaz, M. A., and Vilas-Boas, J. P. (2015). Pad-

dling force proﬁles at different stroke rates in elite

sprint kayaking. Journal of applied biomechanics,

31(4):258–263.

ase, M. (04/2020). Investigation of sensors for force mea-

surement in electrically assisted canoeing. Diploma

thesis, HTW Dresden, Dresden.

Hauptmann, F. (05/2020). Simulation, analysis and control

of the electric drive train of a watercraft. Diploma

thesis, HTW Dresden, Dresden.

Jackson, P. S. (1995). Performance prediction for olympic

kayaks. Journal of sports sciences, 13(3):239–245.

Kaitts Ltd. (2022). Online shop.

Kedean (2022). Online shop.

oschmann, M. (05/2020). Development of a modular en-

ergy storage and management system for LEVs (light

electric vehicles). Diploma thesis, HTW Dresden,

Dresden.

Seidel, J. (02/2022). Analysis of paddle and motion data of

an electriﬁed kayak and optimization of the measure-

ment system. Diploma thesis, HTW Dresden, Dres-

den.

Vikulin, A. (11/2021). Contribution to the modeling and

control of a hybrid driven stand-up paddle. Master

thesis, HTW Dresden, Dresden.

icSPORTS 2022 - 10th International Conference on Sport Sciences Research and Technology Support