Mechatronics Subsystem

The mechatronic subsystem is a tracked chassis designed to transport the robotic manipulator, sensing suite, and payloads through complex environments. The design proposes a platform for precise navigation and manipulation tasks, considering traction, stability under dynamic loads, and energy performance.

Tracked propulsion was chosen over wheeled platforms as a result of superior traction and climbing performance on uneven terrain. A rocker–bogie configuration was considered but rejected due to its higher effective centre of gravity, which is critical during manipulator operation with a heavy payload. The skid-steer configuration also provides high obstacle tolerance whilst enabling zero-radius turning within confined spaces.

Figure A1 – Tracked chassis concept CAD render.

The chassis is driven by two 24V DC motors with integrated planetary gearboxes, transmitting torque through the rear sprockets. Spring-damper elements limit pitch and heave whilst maintaining continuous ground contact. Major masses: battery, motor drives and controllers are packaged centrally between the tracks, minimising centre of gravity height and improving stability during payload handling.

A 24V DC bus powered by a 10Ah Li-ion battery supplies the drive, arm and auxiliary systems. The Simulink power electronics model developed predicts cumulative energy demand within battery limits, enabling 8 full missions per charge [Section D4.3].

 

Figure A2 – Full rover system design in Coppelia.

Figure B1 – Mechatronic subsystem requirements table.

The mechatronic subsystem requirements were derived from the arena constraints, payload specification, and mission profile defined by the system requirements. Key performance drivers include handling of difficult terrain, manoeuvrability within confined spaces, stability under dynamic loads, and sufficient energy capacity for repeated missions. Conservative requirements were set to allow for modelling uncertainty and real-world inefficiencies

 

Figure C1 – Test plan and success matrix.

D1 Geometry and mass envelope

Figure D1.1 – Geometry envelope of chassis.

The rover geometry and mass envelope were defined early in the design process to ensure the compatibility with the operating environment and to constrain subsystem development. All dimensions and mass estimates are derived from a CAD assembly of the chassis, tracks, suspension and drive motors. CAD development was carried out in parallel with simulation to ensure geometric feasibility prior to dynamic verification.

 

 

 

 

 

 

Figure D1.2 – Final chassis geometry.

D1.1 Geometry selection and spatial constraints

The chassis geometry was governed by three primary constraints:

1. terrain traversal (steps, fallen containers, ramps),
2. access through confined indoor spaces,
3. stability during arm operation carrying 12.75kg biohazard

Figure D1.3 – Fallen biohazard clearance.

The highest chassis clearance requirement arises from fallen biohazard containers present within the operating environment.

These containers have a diameter of 0.20m and may be straddled during navigation. A minimum ground clearance of 0.22m was therefore selected to ensure underbody clearance whilst keeping the centre of gravity as low as practicable. Reducing the centre of gravity directly limits both pitch-over during obstacle traversal and roll-over during turning and sloped navigation.

Figure D1.4 – Chassis breakover angle.

A secondary breakover condition was evaluated to assess the risk of high centring over convex terrain features. The breakover angle is defined as

\[ \beta = 2\,\tan^{-1}\!\left(\frac{2h}{L}\right) \]

where h is the minimum chassis ground clearance and L is the effective track ground-contact length (Appendix M2).

For the selected geometry, the resulting breakover angle of approximately 40° exceeds all anticipated terrain features within the operating environment. In addition, the risk of high centring is further reduced by the tracked bogie layout, which distributes contact over multiple wheels and accommodates local height variations.

Front and rear approach/departure angles are maximised by allowing the tracks to lead ahead of the chassis structure, resulting in minimal overhang. Consequently, the dominant geometric limits during obstacle engagement are imposed by suspension geometry rather than by chassis interference.

Figure D1.5 – Lateral clearance in double door constraint and minimum enclosure size.

The chassis footprint constraints were assessed using the environmental dimensions defined in the system requirements (SYS07). The double-door opening provides a total clear width of 3.2m.  This provides 2m of lateral clearance under worst-case orientation.

The tracked configuration allows zero radius turning; hence manoeuvrability is not constrained by wheel steering geometry. As a result, the rover can reorient within the 3 × 3 m minimum closure with sufficient clearance.

D1.2 Track radius and wheelbase

Table E1 summarises the results of the subsystem validation against the defined success criteria. Quantitative performance is extracted from logged simulation data and the Simulink power model, with qualitative behaviour demonstrated via a single integrated test video (Figure E2).

 

 

Figure E1 – Mechatronics subsystem test results matrix.

 

Figure E2 – Integrated mechatronics test video (M01–M07) in CoppeliaSim.

This video demonstrates the slope pull-away and hold, step traversal, envelope navigation, and cruise speed test within a single test arena. The slope hold rollback section is time-compressed for convenience.

Figure E3 – 10% slope drive torque (M01).

Figure E4 – Traction margin at μ = 0.5 (M02).

Figure E5 – Ground clearance during step traversal (M03).

The clearance shortfall occurs during dynamic step interaction, where suspension compression and obstacle approach geometry momentarily reduce clearance. The condition represents an unlikely scenario (step + obstacle coincidence) and does not affect operation on flat or sloped terrain.

Figure E6 – Pitch angle during step negotiation (M04).

Figure E7 – Slope hold rollback over 10 s (M06).

Figure E8 – Cruise and sprint speed performance (M07).

The reported cruise median includes sections of acceleration and deceleration phases. Steady-state tracking aligns closely with the command value of 1.5m/s, suggesting the result reflects inaccuracy in assessment window definition rather than traction.

Figure E9 – Power demand versus battery capability and energy distribution.

Figure E10 – DC bus current versus protection limit.

Figure E11 – Limit flag summary for drive and arm subsystems.

 

 

The tracked 24V chassis concept demonstrably satisfies its requirements, with all criteria passing except ground clearance (M03), validated through full-mission integrated testing.

Electrical performance shows substantial real-world margin: peak bus current (15.68A) and power (376W) remain well below the 20A continuous battery limit, while total mission energy is approximately 12% of usable capacity. This enables 8 complete missions per charge.

Practical deployment constraints were explicitly designed and tested for, including reduced friction, voltage derating, sensor noise, and clearance loss under suspension compression. The resulting margins indicate a robust and conservatively sized system. Traction performance remains within limits at μ = 0.5, below the specified requirement, while stability performance exceeds the design target by over an order of magnitude (θmax = 2.3° vs 32°, rollback = 0.019m vs 0.2m).

Requirement sensitivity highlights clear system boundaries. A tighter pitch constraint (10°) would pass with significant headroom, whereas a steeper slope (20%) would become torque limited, as the 10% launch transient already approaches the motor limit. This is a controlled and well-understood limitation, addressable through launch shaping or increased torque capacity.

The only failure, ground clearance, is a geometric compliance issue rather than a control or power limitation. It can be resolved through modest chassis and suspension revisions. Overall, the subsystem demonstrates strong performance and predictable failure modes.

 

 

Figure R1 – Video of complete Coppelia design navigating test environment.

References:

[1] E. G. Papadopoulos and D. A. Rey, “A new measure of tipover stability margin for mobile manipulators,” in Proc. IEEE Int. Conf. Robot. Autom. (ICRA), Minneapolis, MN, USA, Apr. 1996, pp. 3111–3116, doi: 10.1109/ROBOT.1996.509185.

[2] R. A. Lindemann and C. J. Voorhees, “Mars Exploration Rover mobility assembly design, test and performance,” in Proc. 2005 IEEE Int. Conf. Systems, Man and Cybernetics, Waikoloa, HI, USA, Oct. 2005.

[3] A. Pappalettera, G. Reina, and G. Mantriota, “Design and analysis of tracked stair-climbing robot using innovative suspension system,” Robotics, vol. 13, no. 3, p. 45, 2024, doi: 10.3390/robotics13030045.

[4] Elemex Pty Ltd, “Rectangle 24V 10Ah Lithium-ion Battery Pack with BMS and Bluetooth + Waterproof Case,” Elemex, 2025. [Online]. Available: https://elemex.com.au/product/rectangle-24v-10ah-lithium-ion-battery-pack-with-bms-and-bluetooth-waterproof-case/. [Accessed: Dec. 12, 2026].

[5] Motion Dynamics, “MY1020Z 24V 500W DC Planetary Gear Motor, 500RPM,” product page. [Online]. Available: https://motiondynamics.com.au (accessed Dec. 20, 2026). 

[6] Dimension Engineering, “Sabertooth 2x32 Regenerative Dual Motor Driver,” product datasheet. [Online]. Available: https://www.dimensionengineering.com (accessed Dec. 21, 2026).

[7] StepperOnline, “750W DC Servo Motor Kit, 60V, 2.39 Nm Rated Torque, DSY series,” product page. [Online]. Available: https://www.omc-stepperonline.com (accessed Dec. 21, 2026).

[8] maxon Group, “EC-i 52, Ø52 mm, brushless, 200 W,” motor datasheet. [Online]. Available: https://www.maxongroup.com (accessed Dec. 22, 2026).

[9] Ouster, Inc., “OS1 Digital LiDAR Sensor – Product Datasheet,” rev. OS1-Rev06, 2022. [Online]. Available: https://ouster.com (accessed Dec. 22, 2026).

[10] NVIDIA Corp., “NVIDIA Jetson Orin Nano Series Modules – Data Sheet,” 2024. [Online]. Available: https://developer.nvidia.com/embedded (accessed Dec. 28, 2026).

[11] STMicroelectronics, “VL53L1X Time-of-Flight Distance Sensor,” datasheet. [Online]. Available: https://www.st.com (accessed Dec. 29, 2026).

[12] Bosch Sensortec, “BNO055 Intelligent 9-Axis Absolute Orientation Sensor,” datasheet. [Online]. Available: https://www.bosch-sensortec.com (accessed Dec. 29, 2026). 

 

Appendix:

M1 – Track height
For a circular wheel (or tracked segment) to climb a vertical obstacle unaided, the minimum radius is  
\[ r_{\min}=\frac{h}{2} \]

For the 0.2 m steps (SYS06),  
\[ r_{\min}=\frac{0.2}{2}=0.10~\mathrm{m} \]

A safety factor of 1.5 is applied in line with exploration robot mobility practice [1,2], giving a target effective radius  
\[ r_{\mathrm{des}}=1.5\,r_{\min}=0.15~\mathrm{m} \]

This is met by the selected sprocket and track geometry [1,2].  

M2 – Wheelbase 
Steps are 0.2 m high and at least 0.5 m deep. The diagonal distance from the front edge of one step to the back edge of the next in this minimum case is  
\[ d=\sqrt{0.5^{2}+0.2^{2}}=0.54~\mathrm{m} \]

To maintain at least two stair treads in contact with the tracks and avoid tipping, the chassis wheelbase L is chosen to be at least twice this spacing:  
\[ L=2d=1.08~\mathrm{m} \]

The chosen wheelbase of 1.20 m exceeds this minimum, improving traction and reducing pitch-over risk during stair transitions [2,3].  

M3 – Chassis ground clearance 
The step geometry provides a slope  
\[ \frac{\Delta h}{\Delta x}=\frac{0.2}{0.5}=0.4 \]

which corresponds to an incline of about 21.8°. This represents the worst-case approach and breakover condition. Chassis ground clearance is selected in the main text such that the belly does not contact the step edge at this slope, with margin against manufacturing tolerances and suspension deflection (requirement M03).  

M4 – Drive torque sizing (10% slope, straight climb)  
Inputs (design worst case)  
    Maximum vehicle mass: m=82 kg  
    Gravity: g=9.81 m/s^2  
    Slope: 10 %  \[ \theta=\tan^{-1}(0.1) \]  
    Drive sprocket pitch radius: r_s=0.05 m  
    Rolling resistance (rubber tracks, straight climb): C_rr=0.05  
    Acceleration case: uphill pull-away a=0.5 m/s^2  
    Two tracks share the load equally.  

Required tractive force along the slope  
\[ F_{\mathrm{req}}=mg\left(\sin\theta+C_{rr}\cos\theta\right)+ma \]

Per track,  
\[ F_{\mathrm{track}}=\frac{F_{\mathrm{req}}}{2},\qquad T=F_{\mathrm{track}}\,r_s \]

Computed load cases (per track)  
    Case A – steady/holding uphill (a=0)  
\[ F_{\mathrm{track}}=60~\mathrm{N},\qquad T_A=3.0~\mathrm{N\,m} \]

    Case B – uphill pull-away (a=0.5 m/s^2)  
\[ F_{\mathrm{track}}=80.5~\mathrm{N},\qquad T_B=4.0~\mathrm{N\,m} \]

A design torque requirement per track of  
\[ T_{\mathrm{req}}=4.0~\mathrm{N\,m} \]
is therefore adopted.

Traction check  

With friction coefficient μ=0.6, the available friction force on a 10 % slope is  
\[ F_{\mathrm{fric}}=\mu N=\mu mg=0.6\times 82\times 9.81=483~\mathrm{N} \]

which gives roughly three times margin over the required F_req on a 10 % slope.  

Speed targets  
Using \(\omega=v/r_s\):  
\[ v=1.5~\mathrm{m/s}\Rightarrow \omega=30~\mathrm{rad/s}\Rightarrow 286~\mathrm{rpm} \]
\[ v=2.5~\mathrm{m/s}\Rightarrow \omega=50~\mathrm{rad/s}\Rightarrow 478~\mathrm{rpm} \]

Motor selection  
A MY1020Z 24 V 500 W DC gearmotor with planetary gearbox (6.67:1) is specified per track [5]. Datasheet output torque is 8.5 N·m at about 500 rpm, giving about 2.1× margin over T_req and a theoretical maximum speed  
\[ v_{\max}=2.6~\mathrm{m/s} \]
This satisfies both the slope torque and nominal speed requirements (M06, M07) [5].  

M5 – Track motor electrical demand (per motor)  
Inputs (from M4)  
    Torque cases: τ_A=3.0 N·m, τ_B=4.0 N·m  
    Sprocket radius: r_s=0.05 m  
    Target speed: v=1.5 m/s ⇒ \[ \omega=\frac{v}{r_s}=30~\mathrm{rad/s} \]  
    DC bus voltage: V_bus=24 V  
    Combined efficiency (motor + gearbox + driver): η=0.80  

Mechanical power:  
\[ P_{\mathrm{mech}}=\tau\,\omega \]

    Case A (steady uphill):  
\[ P_{\mathrm{mech}}=3.0\times 30=90~\mathrm{W} \]

    Case B (pull-away):  
\[ P_{\mathrm{mech}}=4.0\times 30=120~\mathrm{W} \]

Electrical power and current:  
\[ P_{\mathrm{elec}}=\frac{P_{\mathrm{mech}}}{\eta},\qquad I=\frac{P_{\mathrm{elec}}}{V_{\mathrm{bus}}} \]

    Case A:  
\[ P_{\mathrm{elec}}=\frac{90}{0.80}=112.5~\mathrm{W},\qquad I=\frac{112.5}{24}=4.69~\mathrm{A} \]

    Case B:  
\[ P_{\mathrm{elec}}=\frac{120}{0.80}=150~\mathrm{W},\qquad I=\frac{150}{24}=6.25~\mathrm{A} \]

Motor datasheet rated current (24 V, 500 W class) is I_rated=26.7 A [5].  The driver selected is a Sabertooth 2×32, rated 32 A continuous and 64 A peak with a 6–30 V nominal input range. For a 6S Li-ion pack the maximum fully charged voltage is  
\[ V_{\max}=6\times 4.2=25.2~\mathrm{V} \]
which is within the driver’s absolute maximum rating.  

M6 – Arm actuator selection (for Simulink power model) 
M6.1 Equations 
Motor torque required for a given joint torque:  
\[ \tau_m=\frac{\tau_{\mathrm{joint}}}{G\,\eta_g} \]

Joint torque capability from a motor and gearbox:  
\[ \tau_{\mathrm{joint,cap}}=\tau_{m,\mathrm{cap}}\,G\,\eta_g \]

Electrical power used in the Simulink model:  
\[ P_{\mathrm{elec}}(t)=\frac{\tau_{\mathrm{joint}}(t)\,\omega_{\mathrm{joint}}(t)}{\eta_g\,\eta_{\mathrm{drv}}} \]
\[ I_{\mathrm{bus}}(t)=\frac{P_{\mathrm{elec}}(t)}{V_{\mathrm{bus}}} \]

Assumptions:  
    Gearbox efficiency: \(\eta_g=0.85\)  
    Servo driver efficiency: \(\eta_{\mathrm{drv}}=0.90\)  
    DC bus voltage: V_bus=24 V  

M6.2 Inputs (Robotics structural sizing)  
Joint static and design torques:  
    Shoulder pitch (J1): τ_static=237.7, τ_design=680.6 N·m  
    Elbow pitch (J2): τ_static=108.9 N·m, τ_design=235.9 N·m  
    Wrist pitch (J3): τ_static=11.5 N·m, τ_design=16.2 N·m  
    Base yaw (J0): τ_design=400.0 N·m  

M6.3 Gear ratio choice and motor selection  
Reference gearbox type: Harmonic Drive CSG-2UH style units with ratios 50:1, 80:1, 100:1, 120:1 and 160:1 and peak torque capacities up to 3419 N·m.  

Motor candidates:  
    StepperOnline DSY 750 W DC servo, rated torque 2.39 N·m, peak torque 7.17 N·m, driver input 24–70 V DC [6].  
    maxon EC-i 52 (24 V, 200 W) BLDC motor, nominal torque 0.646 N·m, stall torque 6.82 N·m [7].  

Sizing check with \(\eta_g=0.85\):  
    Shoulder (J1):  
G=120, τ_(m,rated)=2.39 N·m, τ_(m,peak)=7.17 N·m  
\[ \tau_{\mathrm{joint,cap,rated}}=2.39\times 120\times 0.85=244~\mathrm{N\,m} \]
\[ \tau_{\mathrm{joint,cap,peak}}=7.17\times 120\times 0.85=731~\mathrm{N\,m} \]

    Elbow (J2):  
G=80  
\[ \tau_{\mathrm{joint,cap,rated}}=2.39\times 80\times 0.85=163~\mathrm{N\,m} \]
\[ \tau_{\mathrm{joint,cap,peak}}=7.17\times 80\times 0.85=488~\mathrm{N\,m} \]

    Base yaw (J0):  
G=80, τ_(joint,cap,rated)=163 N·m, τ_(joint,cap,peak)=488 N·m  

    Wrist (J3), maxon EC-i 52:  
G=50, τ_(m,rated)=0.646 N·m, τ_(m,peak)=6.82 N·m  
\[ \tau_{\mathrm{joint,cap,rated}}=0.646\times 50\times 0.85=27.5~\mathrm{N\,m} \]
\[ \tau_{\mathrm{joint,cap,peak}}=6.82\times 50\times 0.85=290~\mathrm{N\,m} \]

These joint torque caps are used as the limits in the Simulink power model [6,7].  

M7 – Auxiliary loads
M7.1 Auxiliary electrical power budget  
Subsystem items and powers used in the Simulink mission energy model:  
    360° LiDAR (e.g. OS1) on 24 V rail: P_LiDAR=16 W  
    Compute (Jetson Orin Nano) on 5 V rail: P_compute=15 W  
    RGB cameras on 5 V rail: P_cameras=4×1.0 W=4 W  
    Proximity ToF sensors (4 units) on 5 V rail: P_ToF=4×0.02 W=0.08 W  
    IMUs (2 units) on 3.3/5 V rail: P_IMU=2×0.04 W=0.08 W  
    Radio / telemetry on 24 V rail: P_comms=7 W (upper bound)  
    DC-DC 24 V → 5 V converter efficiency: \(\eta_{\mathrm{dc,5V}}=0.90\)  

M7.2 Electrical constants used in the Simulink power model
The following constants are implemented in the MATLAB parameter structure and used directly by the Simulink mission energy model:  
    V_bus=24 V (DC bus voltage)  
    η_(drive,total)=0.80 (track motor + gearbox + driver efficiency)  
    η_(arm,total)=0.70 (arm gearbox + servo driver efficiency)  
    η_(dc,5V)=0.90 (24 V → 5 V DC-DC converter efficiency)  
    τ_(drive,limit)=8.5 N·m, I_(drive,limit)=26.7 A (MY1020Z rated torque and current) [5]  
    τ_(arm,yaw,limit)=163 N·m  
    τ_(arm,pitch1,limit)=244 N·m  
    τ_(arm,pitch2,limit)=163 N·m  
    τ_(arm,pitch3,limit)=27.5 N·m  
    (joint torque caps from Robotics structural sizing)  
    I_(arm,limit)=15 A (assumed continuous current per arm joint, used for flagging only)  
    P_(LiDAR,24V)=16 W  
    P_(comms,24V)=7 W  
    P_(5V,total)=19.16 W (total 5 V rail load from auxiliary power budget)  

M8 – CoppeliaSim operation data and power-model input  
The mission energy model is driven by actuator torque and speed logs exported from CoppeliaSim for the full 10-container biohazard recovery mission:  
    Joint torque τ_i(t) and angular velocity ω_i(t) are recorded for:  
        Left and right track drives  
        Base yaw, shoulder, elbow and wrist joints.  
    Logs are exported as a single CSV file at fixed timestep Δt.  

In MATLAB, this CSV is converted into a structured workspace variable and imported into Simulink using a single From Workspace block. This keeps the power model traceable to the validated dynamic simulation.  
The full mission log CSV is attached in the comments.

M9 – Suspension sizing
Worst-case total mass with payload:  
m_total=66.1 kg  

The suspension is approximated as a vertical linear spring–damper supporting the sprung mass  
\[ m_s=m_{\mathrm{total}}-m_{\mathrm{unsprung}}=66.1-9.5=56.6~\mathrm{kg} \]

A static sag target of δ=15 mm is selected to reduce bounce from arm/payload motion while retaining compliance for uneven terrain.  

Total vertical stiffness:  
\[ K_{\mathrm{tot}}=\frac{m_s g}{\delta}=\frac{56.6\times 9.81}{0.015}=3.70\times 10^{4}~\mathrm{N/m} \]

With 8 bogie dampers in parallel:  
\[ k_{\mathrm{each}}=\frac{K_{\mathrm{tot}}}{8}=4.63\times 10^{3}~\mathrm{N/m} \]

Damping ratio target \(\zeta=0.7\) is used for a well-damped second-order response:  
\[ c_{\mathrm{tot}}=2\zeta\sqrt{K_{\mathrm{tot}}\,m_s},\qquad c_{\mathrm{each}}=\frac{c_{\mathrm{tot}}}{8} \]

This gives  
\[ c_{\mathrm{each}}=253~\mathrm{N\,s/m} \]

M10 – Battery pack selection and sizing
The Simulink mission energy model predicts a total electrical energy demand of 21.1 Wh for the full 10-container recovery mission (Section D4.3).  

A 24 V 10 Ah Li-ion e-bike–class traction pack is selected, based on a representative commercial product with datasheet values [4]:  
    Nominal voltage: 24 V  
    Capacity: 10 Ah  
    Nominal energy: 259 Wh  
    Continuous discharge current: 20 A  
    Peak discharge current: 60 A for 3–5 s  
    Pack mass: 1.22 kg  

A conservative usable depth-of-discharge of 70 % is applied:  
\[ E_{\mathrm{usable}}=0.70\times 259=181~\mathrm{Wh} \]

This gives:  
    Missions per charge:  
\[ N_{\mathrm{missions}}=\frac{E_{\mathrm{usable}}}{E_{\mathrm{mission}}}=\frac{181}{21.1}\approx 8\text{–}9 \]

    Simulated peak bus current:  
\[ I_{\mathrm{bus,peak}}=15.7~\mathrm{A}<20~\mathrm{A} \]

So both energy and current requirements are met with margin. The pack mass of 1.22 kg lies within the 4 kg chassis battery allowance. This appendix supports the battery summary in D4.3 (Figure/Table D4.7) and the 20 A current-limit line used in the mission power plots [4].  

Appendix M11 – Tip-over pitch angle calculation 
The forward tip-over pitch angle θ_tip is estimated from the chassis contact length and worst-case centre of gravity height. Static tip-over occurs when the CG line of action passes through the front contact point, giving  
\[ \tan(\theta_{\mathrm{tip}})=\frac{(L/2)}{h_{\mathrm{CG}}} \]

where L is the ground contact length and h_CG is the CG height above ground. For the tracked chassis L=1.2 m and the fully-loaded configuration with the arm fully vertically raised gives h_CG=0.96 m. Substituting yields  
\[ \theta_{\mathrm{tip}}=\tan^{-1}\!\left(\frac{0.6}{0.96}\right)\approx 32^\circ \]

The M04 step-negotiation requirement therefore uses θ_max=32° as the allowable pitch magnitude during ascent and descent.  

Appendix M12 – Full Simulink model pictures: