Preliminary design of a multirotor UAV for indoor search and rescue applications

Multirotor UAVs have become an essential tool in a wider range of applications, including among others disaster management, and search and rescue (SAR) operations. Typically, these systems operate outdoors, with their guidance and positioning being based primarily on GPS. This work is focused on the design and optimization of a multirotor UAV specifically tailored for indoor SAR applications, where GPS signal is unavailable, and obstacles are prevalent. The design incorporates a lightweight frame structure, in order to increase the UAV’s payload capability. This is necessary, since the UAV requires multiple obstacle recognition and avoidance sensors, as well as thermal and optical cameras, to successfully accomplish its mission objectives in a GPS-denied environment. Towards this goal, various trade studies were conducted including different motor/propeller configurations and airframe FEM analyses. The aerodynamic performance of the UAV is evaluated also, using dedicated CFD analyses that incorporate the effect of propellers. Lastly, a prototype of the designed configuration is produced using additive manufacturing methods and initial flight tests of the UAV are performed.


Introduction
Search and rescue (SAR) operations have undergone an evolution with the inclusion of both fixedwing and multirotor Unmanned Aerial Vehicles (UAVs).These technological advancements have revolutionized traditional SAR operations by providing significant advantages in terms of speed, precision, and accessibility [1,2].UAVs equipped with advanced imaging technologies, thermal sensors, and high-resolution cameras can swiftly cover vast terrains, including challenging or hazardous environments that are difficult for human rescuers to navigate.Their ability to capture real-time aerial imagery, identify survivors, and assess disaster-stricken areas aids rescue teams in making informed decisions promptly.Moreover, drones can access confined spaces, such as collapsed buildings or densely vegetated areas, where conventional methods often fail.These capabilities not only enhance the overall efficiency of SAR operations but also significantly reduce response times, crucial in situations where every moment counts.The selection of the UAV to be used depends heavily on the specific requirements of any particular SAR operation (e.g., indoors/outdoors operation, accessibility, take-off and landing area availability).
Specifically for indoor environments, the navigation of UAVs during SAR operations presents a unique set of challenges that demand innovative solutions.The confined spaces, reduced visibility, and intricate layouts within buildings, amplify the complexities of locating and rescuing individuals.One significant challenge lies in the limited access points, hindering both human rescuers and traditional equipment.Additionally, the presence of debris, unstable structures, and potential hazards further jeopardize the safety of rescue teams and survivors alike.Moreover, the urgent need for real-time data in these situations adds pressure, requiring quick and accurate assessments of the environment.To address these challenges, various researchers have proposed innovative solutions.For instance, Petrlik et al. have developed a UAV platform for hazardous GPS-denied areas (i.e.air vents) [3].Sampedro et al. explored the usage of learning-based techniques for a fully autonomous aerial robot for SAR operations [4].These solutions represent significant steps in UAV-aided indoor SAR research, showcasing the diverse and state of the art approaches employed.
In this work, the preliminary design process of an innovative micro quadrotor UAV is presented, which was developed in the framework of the MIDRES project [5].MIDRES is a UAV tailored for indoors GPS-denied SAR operations.It features obstacle avoidance sensors and sophisticated RGBD cameras, enabling it to navigate semi-autonomously in confined, dynamic spaces while simultaneously it has the capability to detect people in distress via a thermal camera.Despite the number of sensors, it is a lightweight, cost-effective 3D printed UAV with compact dimensions, characteristics which are critical as specified during the design requirements.This study delves into the preliminary design of the UAV, covering topics from aerodynamics and structural analyses, as well as trade studies and prototype flight tests with the goal of optimizing each and every component.

Conceptual Design Requirements
The MIDRES project began by administering a targeted questionnaire to its end-users, including entities like the civil protection of the Greek government, firefighting teams, and red-cross teams.The questionnaire carefully addressed 20 pivotal design features / capabilities, critical to the functionality of the UAV.Through these questions the significance of factors such as continuous video streaming, need for a rapid deployment, 10 minutes minimum operating time, extreme temperature conditions adaptability, and size constraint adherence was assessed.The ratings provided by the operators, ranging from 0 (not important) to 5 (very important), assisted the formulation of the UAV requirements (Table 1).These requirements, in conjunction with mission-specific demands, necessitate the incorporation of specialized equipment.Among these, the semi-autonomous navigation system enables the UAV to respond to specific operator commands, such as "move forward 1 meter," allowing autonomous travel between defined points.The operator guides the UAV by providing directional instructions, while the UAV autonomously navigates the route, avoiding obstacles as necessary.This functionality is complemented by live 3D reconstruction of the surrounding environment and obstacle avoidance, facilitated by miniature rangefinder sensors and an RGBD camera ensuring a 360-degree field of view, including upward and downward spaces.Furthermore, meeting the human detection algorithm requirement involves integrating a thermal camera alongside the RGBD camera.The comprehensive list of the required navigation and mission equipment is detailed in Table 2.The conceptual design loop resulted in the prototype depicted in Figure 1, shown both in a CAD representation (Figure 1a) and an actual photo of the integrated UAV.The latter serves as an early prototype, consisting of numerous removable 3D printed parts, which allows rapid modifications, but lacks essential equipment such as sensors, companion computer, and cameras.This platform served as a testing ground to assess the flight controller's capabilities and the adequacy of the propulsion and energy system.Despite its weight being significantly lower than the final version (approximately 40% lighter), valuable observations regarding the maximum thrust-to-weight ratio (T/W) were made.It was noted that the ratio was about 6, with a remarkable 20-minute flight time in a stable hover state.This demonstrated the propulsion system's ability to accommodate additional components with ease, while leaving ample room for optimization through reduction in weight and maximum available thrust, promising enhanced efficiency in subsequent design iterations.

Design Methodology
The design methodology, as depicted in Figure 2, focuses specifically on the preliminary design phase, and employs higher fidelity tools and methods compared to the conceptual design phase.In this stage the creation of a detailed CAD model is desired that will ensure the seamless integration of all essential equipment components.The primary objective during this preliminary design phase is to reduce the UAV's weight without deteriorating its performance and capabilities.This pursuit demands sharp evaluation, since excessive weight reduction might undermine the UAV's structural integrity and longevity.

Frame Design
Fused Filament Fabrication (FFF), the most common 3D printing process [6], was selected as the primary manufacturing method for the UAV frame.This way, an easily customizable and cost-efficient UAV frame can be repeatedly produced within the timespan of a day, without quality compromises.An essential criterion for the frame re-design processes is the limitation that all parts should be accommodated within the confines of a standard commercial 3D printer, like Ultimaker S3.To achieve this, the UAV's main body was strategically divided into three distinct plate compartments.The bottom plate houses the battery, while the middle and upper plates encase critical navigation and payload equipment.Atop of the upper plate, only the companion computer is installed.The UAV's four brackets are securely fastened into position using bolts, allowing for the easy assembly and disassembly.

Structural Analysis
In order to assess the quadrotor UAV's structural integrity, finite element analyses (FEA) were employed in various static conditions, using the ANSYS Mechanical 2021R2 (Academic Multiphysics Campus Solution, Canonsburg, PA, USA) software.The most critical assessment cases, are the ones related to hovering (12000 rpm) and maximum thrust conditions (36000 rpm).Regarding the FEA setup, the 8 main bolts that secure the brackets to the main frame, are grounded (fixed constraints), while in the motor-housing part of the brackets, force, moment and thermal boundary conditions are applied according to the manufacturer's data [7] as presented in Figure 3.While in some cases the effect of the motors' moment and temperature can be considered negligible, in the case of 3D print materials, such as PLA, ABS and PETG, the increased temperature above certain limits will drastically deteriorate their mechanical properties, rendering such a structural analysis unreliable and its results misleading.Hence, it is deemed critical to incorporate the effect of thermal loads in the FEA simulations.

Aerodynamic Analysis
For every major frame re-design, as a result of the previous analyses, dedicated CFD computations were performed to assess the aerodynamic characteristics of the UAV frame, and especially the drag force.
In the initial CFD models only the frame was included, in order to reduce complexity and computational costs.However, in the later stages of the design a fully detailed model of the UAV is investigated, which also incorporates the effect of propeller rotation (Figure 4).The computational grids were produced using the BETA ANSA pre-processing software (v21.0.1, Root, Switzerland), and include both structured and unstructured regions.The computational grid used was the result of a grid dependency study, where the drag force was selected as a monitor variable.The resulting difference in drag between the 2 finer grids was found to be less than 2%.For all grids, a structured zone of 18 cells in the normal to-the-wall direction exists and an appropriate first cell height is selected to ensure that the y + value is maintained everywhere below unity and that at least 5 computational cells lay within the viscous sublayer.The UAV was initially examined in various angles of attack, in order to obtain its drag polar curve.Based on this curve, the correlation between its flight velocity and cruising angle of attack is obtained.Next, to incorporate the effect of propeller rotation, a surrogate model is implemented.More specifically, the actuator disk assumption is employed where mean values for the radial and tangential velocity components are imposed, obtained from dedicated higher fidelity propeller analyses [8] .These velocity values correspond to different propeller rotational speeds which are calculated with the assumption that the UAV performs steady-level flight at a specific cruising angle of attack.

Trade studies results
The preliminary design phase concentrated on reducing the UAV's frame weight, adjusting the propulsion-energy system for optimal performance and autonomy, and identifying the most fitting equipment components tailored to the UAV's compact size.Figure 5 visually illustrates the gradual reduction in both the total UAV weight and the installed equipment weight, encompassing all necessary components, over the course of this specific phase.The starting point is the prototype UAV platform that resulted from conceptual design, namely Dash One, weighing 1840 grams.

Propulsion
The first significant change is related to the propulsion system, leveraging the advantages of a functional flying prototype for rigorous experimentation and refinement.Aiming at optimizing the motor-propeller combination, initial trade studies were conducted, focusing on the T-Motor 2203.5 1500KV motor paired with a HQPROP propeller of 5.1 inches diameter and 2.5 inches geometric pitch.Indoor flight tests provided a controlled environment to explore alternative propeller options while maintaining consistent test conditions.The design space for propellers was centered around the initial choice due to the constrained UAV dimensions.Various rotor sizes, including 4.9, 5, and 5.1 inches in diameter, were tested with different pitch values.The experiment's objective functions were the system's endurance and average electrical current consumption during stable hover flight.Figure 6 demonstrates that a 5.1" × 3.6" propeller yielded extended flight time and minimal electrical current consumption, a critical factor considering the UAV's sensor load.This optimized combination can result in enhanced efficiency and performance in real-world applications.

Equipment
In a parallel approach, similar refinements were applied to other critical equipment components, transcending the choices made during the conceptual design phase.Notably, the initial flight controller (HexCnc Cube 3+) was replaced by the MatekSys H743-v3 model, notable not only for its reduced weight (10 times) but also for its significantly smaller volume.This reduction in volume released valuable space within the UAV, allowing for the integration of additional equipment.Although this model is equipped with a single Inertial Measurement Unit (IMU), to ensure redundancy, the RGBD camera's IMU can serve as a backup in case of the primary's failure.Furthermore, after initial outdoor flight tests, the GPS and compass integrated module, was replaced by a single compact compass unit, hence substantially reducing the module's weight by a factor of five.Additionally, the obstacle avoidance capabilities of the UAV were evaluated, scrutinizing the performance of both the 360° sensors and those oriented upwards and downwards.Leveraging the Ardupilot software in conjunction with the flight controller, the data from all five rangefinder sensors were simultaneously integrated.In order to test the system's capabilities and identify its limits, a challenging real-life obstacle indoor course was devised.Notably, the sensors selected exhibited a robust sample rate, enabling the flight controller to promptly identify obstacles and respond effectively.Given the UAV's forward-focused navigation design, primarily relying on its cameras, the additional sensors functioned as rapid-response mechanisms.When an obstruction was detected, the UAV promptly halted, hovering in place.This pause allowed the operator to assess the surroundings comprehensively before making informed decisions, underscoring the system's efficacy in real-time obstacle management.

Manufacturing
The need for a 3D printed chassis spurred comprehensive trade studies examining various materials based on their structural and thermal integrity, as detailed in Chapter 3. Utilizing Finite Element Method (FEM) analyses, three common materials -Polylactic Acid (PLA), Polyethylene Terephthalate Glycol (PETG), and Acrylonitrile Butadiene Styrene (ABS) -underwent rigorous evaluation according to the material data provided by the manufacturer [9].Key metrics, including maximum equivalent stress and maximum deformation at the UAV's brackets during peak thrust output, were essential for the selection process.PETG emerged as the best choice for this mission, exhibiting the lower attainable maximum stress, while PLA and ABS had a higher maximum stress value by 9.3% and 3.7%, respectively.In terms of thermal resistance, PLA's mechanical properties start to deteriorate at a significantly lower temperature than those of PETG and ABS.During a mission, high temperatures can be experienced due to both external factors, such as ambient temperature rise, and internal, like motors and equipment overheating.Thus, even though PLA has the lowest deformation and the highest yield strength, its low thermal resistance may compromise the whole mission.The above-mentioned findings underscore PETG's suitability as the optimal 3D printing material, ensuring structural integrity, and fulfilling the thermal requirements of the UAV.

Conclusions
In the present study, the design and optimization of a micro quadrotor UAV for indoors SAR applications is presented.Specific design requirements were extracted, and appropriate equipment was selected.Dedicated optimization studies were performed related to structural and aerodynamic aspects as well as improved equipment and propulsion system choices, which resulted in a significant weight reduction and performance enhancement.Three 3D printing materials were examined for their mechanical properties under thermal loads, a parameter crucial for a UAV that its frame is solely 3D printed.Indoor flight tests with simulated real-world scenarios were performed in order to assess the capacity of the obstacle avoidance system and its response rate.The UAV platform that emerged from the preliminary design phase, is illustrated in figure 7, in both a CAD model (7a) and as the real integrated model (7b).MIDRES UAV, equipped with state-of-the-art technological advancements, drives the research in the domain of SAR operations with UAV assistance.

Figure 1a -
Figure 1a -1b.MIDRES prototype in CAD representation (left side) and in an actual photo.

Figure 3 . 4 .
Figure 3. Finite Element Analysis setup Figure 4. CFD grid of the UAV

6 Figure 5 .
Figure 5. UAV and equipment mass reduction during preliminary design loop

Figure 6 .
Figure 6.Flight time and average current consumption during hover with various propellers

Figure 7a -
Figure 7a -7b.Final MIDRES platform in CAD representation (left side) and in an actual photo

Table 1 .
Design requirements of the MIDRES UAV.

Table 2 .
Navigation and mission equipment.