Autonomous Inter-Logistic center transportation service based on Quadcopter

Project Title

Autonomous Inter-Logistic center transportation service based on Quadcopter

Project Area of Specialization RoboticsProject Summary

Unmanned Ariel Vehicle (UAV) based autonomous package delivery (aka drone delivery) is considered one of the innovative services in future smart cities.  Currently  few small scale commercial implementations of such drone delivery services exists like  the Amazon Prime air [2], UAV’s based blood delivery from blood bank to remote hospitals in Rwanda and Ghana in Africa [3],  Vaccination delivery at Melbourne Australia [4] etc.  Despite current commercial deployments, a number of big companies including Google [5], American USP [6],  Walmart [7]  have invested to develop such a service as part of their future portfolio. 

In package delivery via drones, there are two major tasks involved. The package is first attached to the drone (called the pick-up) and is then delivered to the desired location (called the drop-off).  One of the fundamental limitations of current drone delivery services is that only the drop-off part of package delivery is autonomous i.e. the drones are only capable of dropping the package at the desired location. The drones can’t pick-up a package from a desired location as a result of command from the ground station. A drone delivery service with both autonomous pickup and drop-off functionality can enable interesting applications; two of which are summarized in section 1 & 2 (in Benefits of the Project section). Here, in this proposal, we proposed a drone delivery service that provides the capability of both autonomous pick-up of packages and autonomous drop-off. The objective is to develop and test a proof of concept of the proposed service.

1  The proposed solution

Our proof of concept of the proposed autonomous drone delivery service consists of quadcaoptors used as UAVs. Each quadcopter is equipped with GPS and image recognition camera (CMU cam) that help the UAV to arrive at the desired pick-up or drop-off location and precisely land on the landing area. The landing area is in our case is called the docking station. A simple docking station looks similar to a dining table of 2×2 meters in size. At the center of docking station, there is package attachment area. The attachment area is closed until and unless the UAV is centered. The package is inserted in the package box held beneath the attachment area. During the attachment process, the attachment area opened and the package stick upward using a scissor mechanism and is hooked to the quadcopter. Motors and other electronic components of the docking station are controlled using the Raspberry pi onboard PC.

Users of the service are registered through an online web portal hosted on a web server and can order the service via an Android app. The server also hosts other components of the system including statistics collection, request scheduling algorithms, UAV’s route planning etc. Once a package is picked up and delivered, a notification is provided to the user. Meanwhile, the customer can also track the status of their request using the App. 

Project Objectives

The main objective of the project is to build

• A proof of concept drone delivery service with autonomous pick-up and autonomous drop-off capabilities.
• An android and web applications for ordering the service and tracking.
• A docking station with mechanical alignment mechanism to aid in autonomous pick-up and drop-off of packages.
• GPS and Image recognition based precise drone landing on docking station and mechanical alignment to package attachment area.

Project Implementation Method

To achieve the above objectives the project has the following major steps

       1 : Experimenting with SITL UAV simulator? 

Before going to build a quadcopter and test it in the real environment, we will first use the SITL UAV simulator [8].

       2 : Quadcopter building using ArduPilot 

Before going to build a quadcopter and test it in the real environment, we will first use the SITL UAV simulator [9].      

      3 : Route Scheduling through Mission Planner

Mission Planner of AurduPilot software suit will be to schedule the route undertaken by the UAV autonomously, also called a mission. In each mission, single or multiple customer requests can be served before the UAV returned to the depot (starting point). Mission Planner allows different commands to be embedded as of the part of each mission. Each command is an action taken by the UAV, like run the servo motor (to release the package), hover at the specified distance and wait for command from camera module etc. The set of such commands will be used to land the UAV precisely and pick or drop the package at desired location.  

      4 : Docking station implementation

We will implement a simple rectangular docking station made from wood hard board. Beneath the surface of docking station, there will be a package holder where the user can put the package before ordering the service. The center of docking station will be cut into a rectangular shape which we termed as the attachment area. The attachment area will be mechanically opened and closed on demand. All the electronic components of the docking station will be controlled by Raspberry Pi. An example of the docking station is shown in the following figure.

(a) Docking Station Top View	(b) Docking Station Side view

Figure 1: Docking station diagram

          5 : UAV alignment over attachment area using double plotter mechanism

On the top surface of docking station, a mechanical mechanism for aligning quadcoptor to the attachment area will be implemented. Such an alignment mechanism is required because the quadcopter can land at any part (corner) of docking station even if we use GPS and Image processing for precise landing.

            6 : Package Attachment through Scissor Mechanism

Beneath the surface of docking station & bellow the attachment area, a small scissor lift will be implemented. During the pick-up process, the lift is used to push up the package in package holder and attach it to the quadcopter. At the event of drop-off, the lift is used to receive the package. Our scissor mechanism is shown in the following figure.  

Figure 2: Package lifting mechanism

        7 : Development of Web Portal and an  Android App

A website and an android application will be developed to use the proposed drone delivery service. Google map API will be included in the applications to visualize the location of pickup, drop-off or track the current status of an order

(a) Docking Station Top View

Benefits of the Project

  Autonomous drone delivery have several advantages over traditional logistics methods including decrease in delivery and labor cost, reduced delivery time,  low CO2 emission, reduced traffic congestion on roads, improved safety and scalability.  A drone delivery service with both autonomous pickup and drop-off functionality can enable interesting applications; some are discussed in the following paragraph.

Example use cases of proposed solution

1: Proposed Drone delivery service used by CSPs for delivery/pickup directly at customer location

Courier Service Providers (CSPs) like TCS, OCS etc. in Pakistan always want to reduce the package delivery time of their service.  The proposed solution can be used to further reduce the delivery time.   In such a use case, certain customers who require frequent package delivery throughout the day subscribe to the autonomous drone delivery service offered by the CSP. Example of such customers includes the blood collection centers, well-known mobile repairing centers, electronics retailer shops etc. These customers have a docking station installed on the top of the roof of their office premises.

Once a customer wants to send a package to another customer, he/she just put the package in the docking station and opens the Android application in the mobile phone. The customer then select the source and destination of the package and click the submit button. The customer’s mobile phone send the delivery request to the cloud server which then find a free vehicle (drone) to service the request. The selected UAV arrived at the pick-up location and precisely land on the docking station. Once the package is successfully attached the docking station sends an OK message to the cloud server and drone departs to the destination. At the recipient location, the drone once again land on the docking station, is mechanically centered and the attachment is released.

2 Proposed Drone delivery service used between different facilities of a CSP

The use case presented in previous section is equally applicable to transportation service employed between collection offices (e.g. TCS offices) and zonal offices of the same CSP.  Each CSP has multiple collection offices and few zonal offices in big cities where packages from collection offices are collected once or twice a day and are then processed in the zonal office for onward journey. The ever increasing road traffic inside big cities can become a bottleneck in the guaranteed delivery time. The proposed autonomous drone delivery service can avoid such bottleneck. In this use case, the senders of the package delivery are the CSP’s collection centers while the receiver is the zonal office. In contrast to the traditional service, packages can be dispatched to zonal office as soon as received at the collection center (or sufficient packages received to make a bundle for drone delivery). This will expedite the whole delivery process

Technical Details of Final Deliverable

One quadcopter, one or two docking stations, a set of installed ground station software suites and android/web application will be used as a proof of concept. The webportal will be used to register users and provide the details of their respective docking station. 

The android application will be used to order the service.  The user just need to select the recipient of the package as its own location is already known to the system and finally click the submit request button. The request is received at the server that is assigned to one of the UAV. The selected UAV could be either in rest at depot or in flight serving other requests. In the latter case, the route plane of inflight UAV is modified to include the new delivery request before the UAV returned to the depot. Once a package is picked up and delivered, a notification is provided to the user. Meanwhile, the customer can also track the status of their request using the App. 

Final Deliverable of the Project HW/SW integrated systemType of Industry IT , Transportation Technologies Internet of Things (IoT)Sustainable Development Goals Decent Work and Economic Growth, Industry, Innovation and Infrastructure, Sustainable Cities and CommunitiesRequired Resources

Elapsed time in (days or weeks or month or quarter) since start of the project	Milestone	Deliverable
Month 1	understanding the SITL and how to use it	sitl
Month 2	Building Quadcopter with the help of ArduPilot	Quadcopter
Month 3	Exploring Mission Planner (Ground Control Station )	Mission planner
Month 4	Understanding Rapberrypi and controlling stepper Motor	Programming pi
Month 5	Aruco Marker Detection with raspberrypi and a camera	image processing
Month 6	Scissor Lift Building	scissor lift
Month 7	Development of web portal	web site
Month 8	Development of Android App	Android app
Month 9	Final Product	Product
Month 10	Testing	Ok or not