Team Ardra - SEDS VIT

Transcription

1 6f Team Ardra - SEDS VIT VIT University Technical Design Paper for AUVSI Student UAS Competition 2018 Figure 1 Abstract Team Ardra, the UAS team of SEDS-VIT (Students for Exploration and Development of Space) from VIT University consists of 24 talented undergraduate students belonging to various engineering disciplines. This is SEDS VIT s maiden attempt in the competition and the team s main goal for AUVSI SUAS 2018 is to accomplish all the mission tasks successfully. Ardra is FPV model s My Twin Dream, a fixed-wing plane on which the necessary electrical and programming systems have been integrated. This paper provides a description of the developmental progress of Team Ardra in attaining the competition goals in terms of technical headway, systems engineering, safety precautions and economical feasibility. The UAS is capable of executing and completing tasks of autonomous flight, object detection, classification and localization and air delivery. V I T U N I V E R S I T Y S E D S V I T Page 1

2 TABLE OF CONTENTS 1. Systems Engineering Approach Mission Requirement Design Rationale System Design Aircraft Airframe Payload Air Delivery Module Camera Mount Components Placement Propulsion System Autopilot Ground Control Station Interoperatibility Obstacle Avoidance Powering System Imaging System Video Transmission On Board Computers Object Detection, Classification, Localization Autonomous Operation Manual Operations Communications Video Transmission Telemetry Transmission Pilot Communication Antenna Tracker Air Delivery Air Delivery Mechanism in Capsule Air Delivery Drop Time Cyber Security Safety Risks and Mitigation Developmental Risks and Mitigation Mission Risks and Mitigation Conclusion V I T U N I V E R S I T Y S E D S V I T Page 2

3 1. SYSTEMS ENGINEERING APPROACH Systems Engineering is an interdisciplinary approach which introduces the project needs and required functionality early in the development cycle, documenting requirements, then proceeding with design synthesis and system validation while considering the complete problem. Figure 2: The V model 1.1. MISSION REQUIREMENTS In the development process of our UAV Ardra, the V model was implemented, a method which is extensively used in the aeronautical industry to improve the efficiency of the design and production processes. Sequential tasks carried out to meet the design and mission requirements: 1) Work distribution 2) Regular reviews and meetings 3) Integration of all subsystems 4) Testing and evaluation 5) Fail checks and rectification The various tasks to be accomplished according to the SUAS Competition rules were tackled by a thorough analysis of design tradeoffs and building appropriate systems. TIMELINE MISSION TIME TASK STRATEGY DIFFICULTY Avoid taking timeout Primary battery to power propellers, auxiliary battery for all systems mission time becomes 30mins Possible heating up of components and exhaustion of power AUTONOMOUS FLIGHT Using suitable flight controller Optimal path planning OBSTACLE RRT* Algorithm Dynamic obstacles AVOIDANCE *OBJECT DETECTION Off-axis target *ODLC Emergent objects classification Edge detection, template matching and shape, Processing time is high *CHARACTERISTICS character extraction *GEOLOCATION Using time stamp of images and telemetry data Decreasing false positives *AUTONOMY Scripts hierarchy Correct classification, Synchronization between processes AIR DELIVERY Payload drop mechanism which uses links, servo motor and a base plate Wind interference and irresponsive servo motor OPERATIONAL EXCELLENCE Regular meetings, reviews, and interdepartmental communication Lack of previous experience in the competition Table 1: Mission Requirement Analysis V I T U N I V E R S I T Y S E D S V I T Page 3

4 Based on the analysis of the mission requirements, the mission tasks are prioritized in the following manner: PRIORITY High Moderate Low TASK SUBDIVISION Air delivery, Autonomy, Static obstacle avoidance, ODLC Off-axis target Dynamic obstacles Table 2: Inference of analysis The above categorization was made taking into account several factors like Experience and Knowledge, Resources Available for each Task, Complexity of the Task, Average Flight Time, Team Confidence Level, etc DESIGN RATIONALE Team Ardra-SEDS VIT comprises of 24 undergraduates majoring in various engineering departments such as Mechanical, Computer Science, Electrical and Electronics. On the basis of their skillsets and experience, they were divided into the subsystems like Ground Control, Navigation, Payload, Power, Communication, Image Processing and Aircraft. Figure 4: Decision flow process The fixed-wing aircraft My Twin Dream (MTD) was chosen as it is economical, durable, possesses high lift to thrust ratio and also machinable to make modifications as required. The large internal space allowed us to alter the component placement to manipulate the centre of gravity. To assist off-axis object identification, GoPro Hero 5 was chosen due to its small size and low weight. A camera mount was designed to help scan areas that are out of LOS area. MSER is used to find points of interests and Figure 3: System Level Diagram which will then be fed to an algorithm that identifies these characteristics using template matching. The ejection mechanism for the payload drop is placed right on the CG so that the ejection of the bottle won t cause a shift in the CG. The bottle is placed outside the aircraft and is attached with the help of a metal cord so as to cause minimal structural changes to the plane. Due to budget constraints and risk of a crash, a prototype plane was used to test the RRT algorithm for autonomous takeoff, flight and landing. VIT UNIVERSITY SEDS VIT Page 4

5 2. SYSTEM DESIGN 2.1. AIRCRAFT Figure 5: 3D model of MTD airframe in Solidworks Figure 6: Bottom Reinforcements Figure 7: Overall Reinforcements My Twin Dream comprises of four parts: fuselage, two wings, and the tail. The airframe is composed of EPO foam, which is compatible with a wide range of glues, essential for building and repairing of the airplane. Also EPO foam provides glossier surface which is important considering the aerodynamics of the UAV. It is a high wing aircraft whose center of lift is positioned above the center of gravity of the aircraft which makes the UAV more stable and self-correcting during slide-slip, which helps us in achieving the primary task with fewer complications. The airframe has two carbon fiber spar rods running through the wings which add to the overall strength of the aircraft. Through structural analysis, it was found that the fragile parts where stress developed would lead to formation of cracks and its propagation. Hence to avoid that, the parts are covered with foam-tac adhesive making it more durable, increasing the overall flight safety. Also, the wings are removable which makes the transportation of aircraft easier flight stability is achieved by the use 9g servo motors which control the elevator, rudder, and aileron of the aircraft. UAV s cruise velocity is 18 m/s, which is also lesser than the allowable maximum speed of UAV. Tail dragger landing gears are used for safe take-off and landing as they provide more lift to the aircraft during take-off by virtue of the higher angle of attack AIRFRAME Airframe Take-off & Landing Payload Space(cm3) Flight Stability (thrust/weight) Maximum Takeoff Weight(g) Cost($) Spare parts Availability Skywalker X8 Catapult/ Hand Launch 7, Moderate Anaconda (RMRC) Landing Gear(tricycle) 7, Difficult V I T U N I V E R S I T Y S E D S V I T Page 5

6 My Twin Dream Hand Launch/ Landing gear 11, Moderate Table 3: Plane Trade Off After surveying several options, the airframe selected by the team for the 2018 SUAS competition is FPV Model s My Twin Dream, as shown in Figure 1. The My Twin Dream satisfies all of the 2018 AUVSI airframe regulations, by giving large cargo volume and weight and at the same time provides endurance for a long, efficient flight as compared to other notable options as shown in Table 3 and therefore is the selected airframe. Dimensional drawings of the aircraft in Solidworks in metres are shown in Figure 8. Figure 8: (a) Top View (in m) PLANE FEATURES Wingspan Fuselage Length Material SPECIFICATIONS 1800mm 1220mm EPO Foam Take off Hand launch / Runway Landing LiPo Battery Brushless Motors Runway mah 2 x 900Kv Figure 8: (b) Front View (in m) Item Airdrop Mechanism Weight per unit (g) Quantity Total weight (g) Servo Motor Fin Tail Assembly Total 229 Table 5: Air Delivery Module Specifications ESC 2 x 40A Propellers 12 x 6 Thrust Remote 2.8kg (on paper) FlySky TH9X Internal Volume Table 4 : Plane Specification 11,180cm Figure 9: Air Drop Mechanism V I T U N I V E R S I T Y S E D S V I T Page 6

7 PAYLOAD AIR DELIVERY MODULE The air delivery block as shown in Figure 9 has a disfigured T-shape (10 X 60 X 150 mm, bottom part) and (10 X 55 X 150 mm, vertical part). It is made of 3D printed PLA(Poly Lactic Acid) plastic to get a strong, lightweight product. The bottom of the block which is 10 mm thick is fixed to the fuselage. Two protruded blocks are present on the front plane of the air delivery which holds the 9g servo. Further there are two guider blocks with slots of suitable dimensions to constrain the motion of the servo push rod. The design and weight specifications of the Air delivery system have been described in Table CAMERA MOUNT The GoPro camera which is used for object detection, location and classification is placed in a camera mount which is connected to a frame with two base plates fixed to the inside fuselage surface. These base plates provide stability to the frame structure. One side of the frame has the stepper motor which is connected to the GoPro camera mount. The stepper motor rotates from +30 to -30 to cover a wide angle view for the camera. The camera mount is placed at a distance of 588mm from the nose. A hole of 70*40 mm 2 area is made at the bottom part of the fuselage for the lens moment. Fiberglass is attached for aerodynamic and structural safety using epoxy resins. The whole setup ensures that a proper and clean vision for the camera is present and at the same time maintains the stability of the aircraft COMPONENTS PLACEMENTS Figure 10: Camera Mount The auxiliary battery has been placed inside, and the FPV camera has been placed on the nose of the UAS, which is followed by the Pixhawk module. The primary battery is placed near the CG of the UAS and the air delivery module precedes the Odroid. At the rear end of the UAS the camera module is present followed by Ubiquiti system. Figure 11: Components Placement The center of gravity of the aircraft was first calculated manually by calculating the moment generated by each component and then cross verified by assembling these components in the 3D Solidworks model of the UAV. Position of center of gravity was roughly found to be at 2/3 rd of the chord length from the leading edge. To ensure the stability of the UAV while performing the air delivery task, the air delivery module is placed near the CG of the UAV. The same applies with the primary battery, as it is the heaviest part of the UAS. V I T U N I V E R S I T Y S E D S V I T Page 7

8 2.1.3 PROPULSION SYSTEM MTD is a twin electrical motor configuration plane, with one motor on either side of the fuselage. It features a Sunnysky X KV brushless motor set along with a set of Hobbywing 40A brushless ESC to control them. A 4S LiPo mah (20C) battery has been used to power the DC motors through the ESC, Pixhawk and control surfaces. Initially the maximum thrust generated by the cm (9 X4.5 ) diameter propellers, powered by a 4- cell (14.8V) LiPo battery was found out to be 1640g. At later stage of development, due to the addition of few more elements, a larger thrust was required to maintain the thrust to weight ratio. Therefore to provide maximum take off RPM and power consumption of the motors a larger cm (12 X6 ) diameter propeller with a smaller angle of attack which generated 2767g of maximum thrust during static thrust test was used for the aircraft AUTOPILOT An autopilot system controls and maneuvers the trajectory of an aircraft without human intervention. Arduplane, developed as a part of Ardupilot project, is an open source firmware for planes. Ardupilot provides prospect for various autonomous capabilities and hardware interfacing. It is a preferred option and has a wide open source community. It supports various libraries that allow modification depending on the nature of task. Pixhawk PX4 (V ) is selected as it offers a wide variety of features and extensive hardware and firmware support. Pixhawk comes with two integrated advanced processors, Cortex M4 core with FPU 168 MHz and a 32-bit backup failsafe co-processor for manual recovery. It has a redundant power supply, high precision ST Micro 16-bit gyroscope, ST Micro 14-bit accelerometer/compass (magnetometer) and a MEAS barometer. It is compact in size, lightweight and also supports various communication protocols (I2C, SPI, CAN, UART, ADC) which makes it ideal for interfacing with various types of hardwares GROUND CONTROL STATION The GCS consists of Mission Planner and Integrated Control System. Mission Planner is an open source software, which is a community supported application and is officially recommended to be used with Pixhawk PX4 autopilot. It offers more features as compared to other available ground control software such as APM Planner. Mission planner was modified to meet the competition requirements to reflect the obstacles, captures waypoints etc. Integrated Control System (ICS) is responsible for the controlling various other subsystems and for the automation of various tasks associated with the aircraft. It uses various python scripts to perform tasks ranging from Figure 12: Ground Control Station interaction with the Interoperability system to get mission data, upload ODLC and telemetry. It also handles allocation of video feed obtained from the aircraft to systems present under Image Processing System for processing. Autonomous take-off and landing of the plane is also handled by ICS using Mission Planner to control autopilot INTEROPERABILITY The integration with interoperability server is achieved by the ICS using python scripts based on the client library provided for the competition. ICS executes the required python scripts parallel to upload and download a variety of information regarding the mission ranging from waypoint details to ODLC information that is uploaded. V I T U N I V E R S I T Y S E D S V I T Page 8

9 2.3. OBSTACLE AVOIDANCE To accomplish the task of obstacle avoidance, a variant of Rapidlyexploring Random Tree (RRT) path planning algorithm called as RRTstar is employed, which aids in finding the most optimal path a trajectory between the waypoints that minimizes the cost function, subject to given constraints (e.g. obstacles and flight zone). RRT constructs a tree using random sampling of points within the search space. The tree starts from an initial node and expands to find a path towards goal node. The tree gradually expands as the iteration continues. At each iteration, a random node is selected from the sample space. If the random node lies in obstacle free region, then the nearest node is searched in the tree. If it is accessible according to the Figure 13: Node Generation predefined step size, the tree is expanded by connecting the random node and the nearest tree node. Otherwise, it returns a new node by using a steering function, thus expands tree by connecting the new node with the random node. A Boolean collision checking process is performed to ensure the collision-free connection between tree nodes and the random nodes. In addition to this, the nearest neighbor search and rewiring operations are performed to maintain the tree with a minimal cost between tree connections. These processes are performed within the area of a ball of radius given by- A PyGame simulation was first developed to test the capabilities of the algorithm after which it was integrated with the Mission Planner. RRT algorithms are computationally more efficient than similar graph-based algorithms. Hence, the program could run periodically to accomplish the task of avoiding moving obstacles POWERING SYSTEM Taking the calculated power budget into consideration and to comply with the mission requirements, it was decided to use two batteries for the mission. A 4s Lipo mah (20C) battery is used to power the 900KV brushless DC motors through 40A ESC, Pixhawk and control surfaces. A separate 4s Lipo 2800 (35c) mah auxiliary battery is used to power the Odroid XU4, FPV camera, and an ubiquiti antenna. The imagery camera i.e., GoPro Hero 5 comes with its own batteries and has a battery life of around 60 minutes, thus making it suitable for the mission. S. No Component Input Voltage (V) Input Current (A) Power Consumption (W) 1 900kv Motor X W x 2 = W (at max throttle) 2 12g Servos X W x 4 = 6W 3 Pixhawk PX W 4 GPS Module W 5 Rfd W 6 Air Speed Sensor W Power distribution board Matek 6s is used to provide dedicated working voltages to both the motors. It also has 5V and 10V BECs for providing regulated Total A W and filtered output power. It has a maximum input voltage Table 6: Primary Battery Power Calculation and filtered output power. It has a maximum input voltage of 27V, thus making it suitable for 4s Lipo battery. Pixhawk power module has been used to power the Pixhawk which provided a stable 5.37V and 2.25A power supply to reduce the chances of burn out. V I T U N I V E R S I T Y S E D S V I T Page 9

10 S. No 2.5 IMAGING SYSTEM Component Input Voltage (V) Input Current (A) Table 7: Auxiliary Battery Power Calculation Power Consumption(W) 1 Odroid XU W 2 Ubiquiti Loco M W 3 FPV Camera W Total 5.5 A 37.3 W The system has two types of the camera systems. A primary camera for Object Detection, Classification & Localization and an FPV (First Person View) camera for manual takeover in case of any emergency situation. GoPro Hero 5 black is chosen as the primary camera due to its compact size, lightweight, self- stabilization and a capability to support 4K videos. The plane would be flying at a height of 150 feet and the video from the primary camera will be sent at 1920 x 1080 resolution, since the object is of size 1 foot, it will take up 15 x 15 pixels of the whole image. The primary camera is mounted on a custom-made gimbal which allows horizontal rotation to capture off-axis targets. For the FPV camera, Logitech C270 HD is chosen which provides high-resolution video feed (1280 x 720) and thus ensuring easy manual control. Both of the cameras are controlled by the onboard processor. Camera Price ($) Weight (gm) Maximum Resolution Dimensions(cm) Nikon D x x 7.5 x 9.8 Canon PowerShot SX540HS x x 22.9 x 15.6 GoPro Hero 5 Black x 6.2 x 4.5 Table 8: Imaging Camera Tradeoff VIDEO TRANSMISSION The real time video from the FPV camera is live streamed to the ground station in case of manual takeover with the help of the onboard processor (Odroid xu4) at a moderate video quality (480x360) to avoid too much load on the processor and the antenna. V I T U N I V E R S I T Y S E D S V I T Page 10

11 The video from the primary camera is saved on the onboard processor first, at high quality (1920x1080) and then transmitted to the ground station. This is done so that a backup of the video is always present even if there is a connection loss. The saved files can be retransmitted once connection is re-established ON-BOARD COMPUTERS OBC Clock Speed RAM (GB) Size (mm) GPU Weight ( Kg) Raspberry Pi 3 Model B 1.2 GHz 1 GB 122 x 76 x 34 Broadcom video core Odroid XU4 2 GHz 2 GB 83 x 58 x 22 Mali-T628 MP Nvidia Jetson TX1 1.9 GHz 4 GB Core NVIDIA 1.54 Table 9: On-Board Computer Tradeoff Odroid XU4 is selected as the OBC as it far outweighs other computers and is best overall choice. It has optimal dimensions and weight making it ideal for use on an UAV. Further it has a considerable amount of dynamic memory and good clock speed making it fast and allows it to multitask efficiently OBJECT DETECTION, CLASSIFICATION, LOCALIZATION AUTONOMOUS OPERATIONS The Autonomous Detection, Localization and Classification system comprises of several subsystems to find and characterize static and emergent objects successfully. At first, the video is captured by the primary camera (GoPro Hero 5 Black) and sent to the GCS by the onboard processor through a 5.8 GHz channel. The frames are then pre-processed to identify the regions of interest. This is done by the use of Maximally Stable External Regions (MSER) feature descriptor. The output at this stage consists of cropped images of the probable targets. The obtained cropped images are segmented to separate the shape and the character. Canny edge detection is applied to find the edges following which contours are detected. The contour with the highest area is supposed to represent the object, using which the character s contours are detected. The shape and character are separated for classification. The shape is classified using the template matching technique. The matchshapes function is used to find the degree of similarity between the contours of the shape and the template. The template corresponding to the lowest value is selected as the shape. V I T U N I V E R S I T Y S E D S V I T Page 11

12 For character detection, templates of the characters are taken and rotated in 8 different directions. The angle at which a character class gets the highest score is then selected as the shape orientation and the corresponding class is taken as the character. Figure 14: Autonomous Imaging System Color classification is done by masking the cropped image with all the known colors. For each color, the areas of all regions containing the color are calculated. The color which produces the region with the highest area is assigned to the shape and the one with next highest area is assigned to the character MANUAL OPERATIONS If a target is found and the algorithm cannot identify the specific properties of the target, the cropped region of interest will be saved to the file and the properties will be specified by the operator on the ground COMMUNICATIONS Figure 15: Communication Systems The Communication subsystem consists of three channels namely, 915MHz (Telemetry, Transmission and Reception), 2.4GHz (Manual Takeover) and 5.8GHz (Imaging Transmission). V I T U N I V E R S I T Y S E D S V I T Page 12

13 VIDEO TRANSMISSION Model Transfer Speed Range Dimensions Weight (Kg) Ubiquiti PBE-5AC Mbps 25+ km 520 x 520 x 308 mm 2.35 Ubiquiti NBE-5AC Mbps 15+ km 140 x 140 x 54 mm Ubiquiti NSM5 150+Mbps 15+ km 294 x 31 x 80 mm Ubiquiti LOCOM5 150+Mbps 10+ km 161 x 31 x 80 mm Table 10: Imaging Antenna Tradeoff Ubiquiti loco M5 is selected out of all available choices. It is chosen primarily due to its light weight and small size. Also it has a considerable range and transfer speed which fulfills our mission requirements TELEMETRY TRANSMISSION RFD 900+ is primarily used for telemetry communication and sending waypoints and receiving status messages from the autopilot. It offers good data rate and range which ensures continuous communication PILOT COMMUNICATION Transmitter (FlySky FS-TH9X) is used to control the plane if autopilot ceases to function. It acts as failsafe in case the system malfunctions. The on-board receiver (FS-R9B) takes over the controls if the override is activated. Transmitter and receiver work on 2.4 GHz band using AFHDS communication protocol, a digital protocol that ensures 2 or more radios can operate at the same time without interfering with each other ANTENNA TRACKER The antenna tracker employs an Ubiquiti Loco M5 and RFD900+ antenna. It tracks the movements of the plane and a robust mechanism moving in the direction of the plane to avoid any loss in data packets. It follows a two degree of freedom motion. The base rotates from -180 degrees to 180 degrees and covers the x-y axis and is responsible for the Pan movement. Secondly, the tilt motion makes sure the tracker matches the altitude of the plane and moves in sync with the plane s motion. The second degree is restricted from 0 to 90 degrees. Figure 16: Antenna Tracker V I T U N I V E R S I T Y S E D S V I T Page 13

14 The tracker uses a simple calculation to get the required angles. The three coordinates of the planes are used to calculate the angles of the tracker in 3D space AIR DELIVERY AIR DELIVERY MECHANISM & CAPSULE At zero position of servo motor, the bottle is fixed to the fin-tail assembly which hangs on the torque rod, outside the aircraft, by means of a steel cord. After calculating the optimal time to dispatch the bottle, upon deployment the servo motor arm rotates. This motion pulls back the torque rod from the right guider block to the left one, which releases the air delivery capsule. Figure 17: Air Delivery Mechanism To ensure that the water bottle breaks open upon impact, a conical fin-tail assembly is used which will ensure that the bottle hits the ground, nose heading down. The conical fin-tail assembly will increase the stability of the bottle during its free fall. Also, it will help us to hit the target with greater accuracy. The fin-tail assembly was manufactured using 3D printed PLA plastic and designed taking a research paper as reference AIR DELIVERY DROP TIME An automated method was identified to calculate the optimal time to release the bottle. The optimal time depends on the velocity and altitude of the plane, distance between the drop and UAS location. Velocity of the aircraft is measured using the pitot tube placed on the nose of the UAS, altitude is calculated by Pixhawk, and distance is calculated by the GPS coordinates of the UAS. Assuming the bottle follows a parabolic path with no considerable drag, the optimal time to release the bottle can be easily calculated by solving the equation of projectile motion with help of the above mentioned variables CYBER SECURITY Radio link protection : The telemetry data of the plane is sent through using RFD900+ which uses Advanced Encryption Standard (AES) and frequency hopping to secure the radio communication link between the aircraft and GCS. GPS: The GPS module used on the plane i.e. Ublox Neo 7m comes with inbuilt anti-jamming mechanisms which uses Active CW (Continuous Wave) jamming detection and removal algorithm and an extra on board SAW band pass filter to attenuate out-of-band signals to prevent jamming. Wi-Fi link protection: The communication between the plane and the GCS for transmission of FPV feed and object images is done through Wi-Fi link using Ubiquity Nanostation Loco M5. The link is protected using the WPA2 security protocol. Further to avoid signal jamming, Nanostation automatically switches to lowest noise frequency in 5.0 GHz to 5.8 GHz band. V I T U N I V E R S I T Y S E D S V I T Page 14

15 3. SAFETY, RISKS AND MITIGATION STRATEGIES Safety and precaution methods are extremely vital to the proper functioning and operation of an Unmanned Aerial System. The various risks faced/ most likely to occur during the development process and mission are listed and the mitigation measures taken to prevent them are tabulated below DEVELOPMENTAL RISKS & MITIGATION STRATEGIES S.NO RISKS MITIGATION MEASURES 1 Injury due to bad usage of tools Safety measures and precautions followed in the lab. Equipments handled with care 2 Electrical malfunction/ damage Proper taping of wires with Si glue and back-up components 3 Increase in weight of Aircraft High thrust-weight ratio provided using 12 x 6 propellers 4 Mistakes by the pilot Training and practice sessions for the pilot 5 Breakage of frame Reinforcement of critical parts using foam-tac to prevent crack formation 6 Li-Po battery fire Foam covering to prevent explosion in case of crash 7 Not enough funds to finance the project Fundraiser / Crowd-funding link circulated on Social media pages 3.2. MISSION RISKS & MITIGATION STRATEGIES S.NO RISKS MITIGATION MEASURES 1 Loss of control over the plane Pixhawk will automatically switch to the RTL(Return to Land) mode 2 Failure to meet mission time limits 2 batteries to power communication and propellers separately, hence calculated flight time is 30mins 3 Unexpected payload drop Payload is secured safely using a link 4 Shifting of Center of Gravity of plane due to dislodging of components Components are secured internally using adhesives like Cyno 5 Air Delivery Mechanism (ADM) Instantaneous manual override to drop the payload Malfunction 6 Loss of WiFi Link Presence of onboard processor, hence data can be retrieved later from the plane 7 Loss of Communication Autopilot will override the manual system and control the plane 4. CONCLUSION Our journey for SUAS 2018 has been a long and steady one with numerous obstacles and difficulties along the way. This being our first year in this competition we gained knowledge and experience regarding the working, manufacturing and characteristics of an UAS which helped us refine its performance and durability for long term benefits with respect to the main mission objectives. Using the systems engineering approach we looked at aspects of safety, endurance and reliability for selecting tasks for us to focus on. With regular and extensive testing of individual components and the MTD as a whole, the team has developed an unmanned aerial system which will be capable of accomplishing the required set of mission tasks in the competition. V I T U N I V E R S I T Y S E D S V I T Page 15