Assisted Living Robotic Design Concept

 

Assisted Living Robotic Design Concept

by

Stanley D. Pebsworth

Embry-Riddle Aeronautical University

December 2015

                                   

A Design Research Project submitted to the Worldwide Campus in partial fulfillment of the requirements for course ASCI 531, Robotics and Control

Abstract

This design project focuses on the application of assisted living robotics.  In an era of advanced robotic design, assisted living robots are within the grasp of humans with special needs.  This project will focus on the current available systems and provide applications for those systems.  It will also review current assisted living issues that could be met through the use of robotic assisted living butlers.  In the future, demands for assisted living needs will be higher and the availability of nurses will be lower.  Through the use of assisted living robotics, we can bridge this gap.  Many people feel the need to remain independent; this design project will review scholarly research the addresses the possible acceptance of these systems.  With the invent of new technology, the cost of robotic assisted living can be reduced.  This project will attempt to project the future costs of robotic assisted living butlers.  There are many misconceptions and negative attitudes towards robots however, this paper will show that many are enthusiastic and optimistic about the possibilities that assisted living robotics provide as well as highlight the need for continued research to better understand the support that robots can provide.  Even though it may seem much like a futuristic movie, assisted living robots will someday be a part of our everyday lives.

Keywords: robotics, automation, humanoid, butler, assisted living

Introduction

            A 1998 publication by HONDA that presented its humanoid robot the P2, paved the way for future research and development in human-like robots dedicated to servicing humans.  Giving robots human characteristics and behavior is just the beginning.  It is believed that giving a robot a more human-like appearance tricks our subconscious and helps us look beyond the robotic-like characteristics and behaviors and see a more complex object (Tondu, 2012).

            There have been several human-like robotic design up to now.  Currently, this research has become a very exciting topic.  The success of HONDA’s P2 is believed to have triggered the worldwide research of human-like robots.  Up to this point in the design and research of these systems, there has still been little achieved in application of these systems.  A new promising advancement was however revealed in 2004 called Advanced Step in Innovative Mobility (ASIMO).  This new technology allows the robot to move, walk, and run much like a human (Kaneko, Kanehiro, Morisawa, Tsuji, Miura, Nakaoka, & Yokoi, 2011).

            Future needs are shifting from industrial manufacturing robotics to a more human-like robotic system.  Our surroundings were designed for human needs and therefore the need for a more human-sized robot will reduce costs to adapt the robotic systems to our current work environments.  The need for these systems has generated a need for research and development that is currently being carried out by private companies and universities worldwide.  The most impressive humanoid robots to date are the HONDA P2, ASIMO, HUBO2, LOLA, BHR-2, iCub, Lucy, REEM-B, as well as the HRP platforms (Kaneko, Kanehiro, Morisawa, Akachi, Miyamori, Hayashi, & Kanehira, 2011).

            In order for these human service robots to become a reality for our future we face several technological issues.  Production costs are one of these issues.  These production costs have been reduced for small humanoid systems that are less than 50cm however, for human-sized robots, this reduction in cost is not foreseeable in the near future due to research and development costs.  Another issue is power.  In order for these systems to coexist with humans, they will need to have power sources that allow them to function for several hours at a time.  The HPR-4 has developed technology that allows it to use 30% less power than previous models (Kaneko, et.al. 2011).

            At the current pace of this technology.  The researcher believes that we will see assisted living style robots in our near future.  These systems have the potential to provide mobility for those that have lost it due to accidents or disease and improve their quality of life.  Private companies and universities must develop ways to reduce cost for these systems as well as develop ways for these human-like robots to use as little power as possible.  The HRP-4 project is very likely to be the person you ask to get your next cup of coffee and hand you your morning paper.

Design Concept

            Based on the current pace of the humanoid robotic industry, this researcher believes the best design for an assisted living robot will be one that has a human-like appearance.  The design concepts that will be addressed in this design are as follows.

            1) Lightweight body comparable to the average human

            2) Price point must be attainable

            3) Power consumption must provide an eight hour run time

            4) Object manipulation must dexterity must be human-like

            5) Must be upgradable

Design Decisions

            Lightweight body design was considered in order to limit the requirement for transportation and home modifications.  By creating a robot that is close in stature and weight to a human there should be no need to modify current vehicles or homes where these systems are required to operate in.  This in turn will help to lower the price point of the system due to there being no need for external equipment to move, manipulate, house, or store this system.  Price point will also be addressed by requiring modular systems in the design.

            Power consumption will be addressed to allow for up to eight hours of operational time since these robots will be working closely with humans and a typical eight-hour work day was considered.  As power is addressed, close attention will be paid to the output power of motors used in the design both for power consumption reason as well as possible safety issues as they will be working closely with humans.

            Object manipulation and upgradability will be considered and based on current research, 7-DOF arms will be utilized to provide more human-like movement as well as improved dexterity.  To also aide in dexterity, the design will consider the ability to upgrade software as enhancements are made.  This upgradability will be addressed in the design by providing networkability of the hardware that will allow the designer/developer to upgrade the system remotely as well as download operational parameters to ensure proper operation of the system.

Specifications

            In order to give this design a human-like appearance and stature based on the previous design concepts, the basic specifications considered will be those in Table 1.  The outward appearance of this design will be similar to the HRP-4C as seen in Figure 1.

Table 1. Basic Design Specification

Dimension	Height	66in
	Width	25in
	Depth	13in
Weight with Batteries		110lbs
D.O.F.	Head	2 DOF
	Arm	2 x 7 DOF
	Hand	2 x 2 DOF
	Waist	2 DOF
	Leg	2 x 6 DOF
	Total	34 DOF
Control System		Distributed System
Batteries	Type	NiMH
	Specification	DC 48V @ 5.4Ah

Movement, Control, and Sensing

            Walking control can be a complex problem for biped robots due to the large number of constraints and no fixed reference frame.  In order to provide stable motion, reaction forces have to be controlled indirectly by the motion of the whole system.  In order to achieve the desired movement stability of this design we will consider computations for control as outlined by Santacruz, C., & Nakamura, Y. (2012).  Santacruz & Nakamura propose the use of minimal energy control to generate the desired trajectory of the system.  With this method of stabilization, the design concept will walk and follow a desired pattern as specified by a motion pattern as well as a force pattern.

            Aiding walking will be the incorporation of sampling-based trajectory imitation also known as Laplacian-RRT.  This algorithm will aide this design concept as it maneuvers in constrained environments similar to a home environment.  This concept has been proven in simulation as well as in HRP-4 robot and allows for increased convergence speed (Nierhoff, T., Hirche, S., & Nakamura, Y., 2014).  Information relied on for this algorithm will be supplied using 3D vision.  3D vision will also assist in the design concept of vision guided homing that will allow the design concept to incorporate hand eye coordinated movements as proven and tested on the HARO-1 humanoid service robot (Jin,Y., & Xie, M., 2000).

            New generation tactile modules will be considered for this design concept in order to incorporate a whole body touch sensation.  These modules will allow this robotic design to better interact with its environment.  A local processor will be incorporated that will eliminate the need for networking and high processing bandwidth requirements.  The modules considered have been proven and tested by Mittendorfer, P., & Cheng, G. (2011).  Mittendorfer, P., & Cheng, G. (2011), state that the sense of touch is a large portion of our proprioceptive system and assists us in planning tasks and motion control.  The design concept considered in this paper will benefit greatly from the use of tactile-sensing skin.  Limiting this design to simple sensing modules would create large errors in sensing object geometry and location (Dollar, A., Jentoft, L., Gao, J., & Howe, R., 2009).

            This design concept will also consider the research of Chen, X., & Yangang, W. (2013), regarding knowledge management systems for humanoid robots.  Their design concept is based on a design catalogue and is divided as follows.

            1) Object Catalogue comprising of specific task like movement.

            2) Operation Catalogue comprised of the working principal of the program design

            3) Solution Catalogue comprised of the solutions for specific tasks

Figure 1. The structure of design information catalogue of the humanoid robot knowledge base. Chen, X., & Yangang, W. (2013)

Logic Design

            Humanoid robots pose an intriguing shift in our acceptance of robotic systems that we interact with on a daily basis.  By incorporating emotion models and expressive behavior into robotic design we increase the acceptance of these systems in our environment.  Traditionally, autonomous robots were not designed to nor any considerations made, for them to interact with humans.  Theorist argue that simple expressions such as disgust, anger, fear, and joy serve particular biological or social function.  It has been scientifically proven that robots that closely resemble humans in look and expression are more readily accepted as partners rather than tools (Brezeal, C., 2003).  These anthropomorphic projections are crucial in the acceptance or rejection of the human perception that a robot is perceived to be human-like (Tondu, B., 2012). Therefore, it is the intent of this design concept to be as sociably human-like as possible.  Figure 2 proposes that the visual design concept intended for this project be similar to that of the HRP-4C.

Figure 2. Front views of HRP-3 (Left), HRP-4C (2nd Left), Body Mechanism of HRP-4C (3rd Left), HRP-4 (Middle), Side views of HRP-4C (3rd Right), HRP-2 (2nd Right), and front view of HRP-2 (Right). Kaneko, K., Kanehiro, F., Morisawa, M., Akachi, K., Miyamori, G., Hayashi, A., & Kanehira, N. (2011)

Conclusion

            This paper presents the design concept of a humanoid assisted living robot.  Design concepts used have been proven through science and or research.  It is the intent of this paper to propose a design that would be easily accepted to work in close proximity with patients, the handicapped, and or those simply requiring assisted living.  It is also the intent of this paper to propose a theory of design and not imply that this design in its completion is a tested design.  This researcher feels that with a collaborated effort, this design concept could be a viable future humanoid robotic system that would easily adapt to an assisted living roll.

References

Brezeal, C. (2003). Emotion and sociable humanoid robots. International Journal of Human - Computer Studies, 59(1), 119-155. doi:10.1016/S1071-5819(03)00018-1

Chen, X., & Yangang, W. (2013). Knowledge management system for humanoid robot based on design catalogue. International Journal of Advancements in Computing Technology, 5(9), 1183.

Dollar, A., Jentoft, L., Gao, J., & Howe, R. (2009). Contact Sensing and grasping performance of compliant hands. Autonomous Robots, 28, 65-75. doi:10.1007/s10514-009-9144-9

Ghidoni, S., Anzalone, S. M., Munaro, M., Michieletto, S., & Menegatti, E. (2014). A distributed perception infrastructure for robot assisted living. Robotics and Autonomous Systems, 62(9), 1316-1328. doi:10.1016/j.robot.2014.03.022

Goeldner, M., Herstatt, C., & Tietze, F. (2015). The emergence of care robotics - A patent and publication analysis. Technological Forecasting and Social Change, 92, 115-131. doi:10.1016/j.techfore.2014.09.005

Jin,Y., & Xie, M. (2000). Vision guided homing for humanoid service robot. Paper presented at the School of MPE, Nanyang Technological University, 4 511-514 vol.4. doi:10.1109/ICPR.2000.902969

Kaeko, K., Kanehiro, F., Morisawa, M., Akachi, K., Miyamori, G., Hayashi, A., & Kanehira, N. (2011). Humanoid robot HRP-4 - humanoid robotics platform with lightweight and slim body. Paper presented at the 4400-4407. doi:10.1109/IROS.2011.6094465

 Kaneko, K., Kanehiro, F., Morisawa, M., Tsuji, T., Miura, K., Nakaoka, S., Yokoi, K. (2011). Hardware improvement of cybernetic human HRP-4C for entertainment use. Paper presented at the 4392-4399. doi:10.1109/IROS.2011.6094415

Kawanda Industries Inc. (n.d.). Humanoid Robot HRP-4. Retrieved 18 November 2015 from http://global.kawada.jp/mechatronics/hrp4.html

Kutner, N. G., Zhang, R., Butler, A. J., Wolf, S. L., & Alberts, J. L. (2010). Quality-of-life change associated with robotic-assisted therapy to improve hand motor function in patients with subacute stroke: A randomized clinical trial. Physical Therapy, 90(4), 493-504. doi:10.2522/ptj.20090160

Nierhoff, T., Hirche, S., & Nakamura, Y. (2014). Sampling-based trajectory imitation in constrained environments using laplacian-RRT. Paper presented at the 3012-3018. doi:10.1109/IROS.2014.6942978

Mittendorfer, P., & Cheng, G. (2011). Humanoid multimodal tactile-sensing modules. IEEE Transactions on Robotics, 27(3), 401-410. doi:10.1109/TRO.2011.2106330

Santacruz, C., & Nakamura, Y. (2012). Analytical real-time pattern generation for trajectory modification and footstep replanning of humanoid robots. Paper presented at the 2095-2100. doi:10.1109/IROS.2012.6386216

Spenko, M., Yu, H., & Dubowsky, S. (2006). Robotic personal aids for mobility and monitoring for the elderly. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 14(3), 344-351. doi:10.1109/TNSRE.2006.881534

Tondu, B. (2012). Anthropomorphism and service humanoid robots: An ambiguous relationship. Industrial Robot: An International Journal, 39(6), 609-618. doi:10.1108/01439911211268840

Automated Unmanned Refuse Collection (AURC)

 

Automated Unmanned Refuse Collection (AURC)

by

Stanley D. Pebsworth

Embry-Riddle Aeronautical University

October 2015

                                   

An Application Project submitted to the Worldwide Campus in partial fulfillment of the requirements for course UNSY 501, Application of Unmanned Systems

Abstract

This application project aims to address the idea of unmanned refuse collection.  Refuse collection is a process that requires coordinated efforts to both timely collect and properly transport refuse to a disposal site.  These coordinated efforts, if not done properly, result in late routes, personnel overtime, as well as safety concerns.  This project will assess the need for Automated Unmanned Refuse Collection (AURC) and evaluate the implications of its use.  This project will address the technological, social, environmental, and political ramifications of AURC.  It will also address contributing factors and effects that will show just cause for this solution.  Legal, ethical, and safety concerns will be addressed and a specific research strategy will be proposed.

Keywords: automation, unmanned systems, refuse collection, waste management

Automated Unmanned Refuse Collection

            Routing of refuse collection to allow for scheduled and timely delivery back to the landfill is a timely and costly effort.  The human interface in these systems results in missed routes, late departures and arrivals, as well as congestion during refuse drop off.  There is also the inherent risk involved in refuse collection.  By removing the human factor in refuse collection, we can better route collection vehicles, coordinate departures and arrivals, practically eliminate congestion during refuse drop off and reduce safety related accidents in the workplace.

            It is time for a change in the current model of refuse collection.  For many years the collection process has gone relatively unchanged.  Cities are growing and the amount of refuse grows with them.  Money is being thrown away due to poor route management and non-optimization of vehicles on collection routes.  This must come to a stop.  We must improve routing and optimize the use of collection vehicles through automated systems.

Problem Statement

            So how can we improve on our current refuse collection procedures?  What methods are available to address our issues?  Several methods have been proposed to combat these issues.  The most popular of which is the use of  automated unmanned refuse collection vehicles.  These trucks can be routed and controlled though a central facility that will manage all aspects of the refuse collection and disposal.  Other methods have included simply automating routing and maintaining the manned collection vehicles.  This method will not remove the very cause of our issues; the human interface.

            We must look for alternatives that will be safe, save time, and provide customers with better service.  The problems we face today in refuse collection is scheduling of refuse collection vehicles as efficiently as possible.  When a refuse collection truck returns half full, it costs the same amount of money to your company as if it were completely full.  So how do we address this issue?  How do we make the most of every trip our refuse collection vehicles make?

Background

            Refuse management is a mandatory and essential service provided by municipal authorities in order to keep our streets clean.  Today’s systems for refuse collection are unscientific, outdated and inefficient.  As our populations grow, we are faced with managing the challenges of efficiency, increased rates of refuse generation, high collection costs and fewer financial resources (Huang, Lin, 2015).

            Seadon (2010) says that in order to move to a more sustainable society, it is required that we develop greater sophistication to manage waste.  He also states that our traditional reductionist approach is not sustainable because it has no flexibility or future development.  We must develop a waste collection system that is more adaptable and has plans for future development.

     Waste is a result of inadequate thinking. The traditional approaches to waste management of “flame, flush or fling” are outmoded customs which have resulted in an unsustainable society. In the USA the total annual wastes exceed 115 billion tons, of which 80% is wastewater. Of that amount less than 2% is recycled. Emitting waste into the environment resulted in nearly 40% of all USA waters being too polluted to support their designated functions and more than 45% of the USA population live in areas where air quality was unhealthy at times because of high levels of air pollutants (Seadon, 2010).

In 2010 garbage collection was rated as the seventh most dangerous job in the United States.  The mortality rate for waste collection workers in 2010 was 3 in 10,000.  In Canada, a 2010 report stated that garbage collection was “one of the most hazardous jobs.”  It was also stated that injuries had affected 35% of workers for that same year in Canada.  In order to combat these safety issues, companies normally require refuse collection workers to wear leather gloves, long pants and heavy steel-toed boots (Tibbetts, 2013).

Waste collection has contributed greatly to human health issues in the United States as well.  In a 1998 report from the United States Bureau of Labor and Statistics reported that U.S. waste collectors died due to work related issues at a rate of 49 per 100,000 workers and nonfatal accidents were at a rate of 95 per 1,000.  This report also ranked U.S. waste collection as the seventh most hazardous job in the U.S.  The most identifiable hazards to waste collectors were items such as dust, endotoxins, bacteria and fungi.  These hazards resulted in acute issues such as respiratory symptoms, gastrointestinal symptoms, hepatitis, HIV, and syphilis.  (Kuijer, P., Frings-sen, M., & Sluiter, J. (2010).

Kuijer, P. et al. (2010) further reported that in a 2008 study, nearly 38% of waste collection workers reported an on the job injury.  The causes of these injuries happened mainly away from the collection facility.  Workers were hit by goods or a vehicle, fell from a high elevation or stepped on or fell on a sharp object.  When waste is collected manually, specific consideration should be taken as to the increased risk for hearing loss, respiratory issues, gastrointestinal problems, and back injuries.

            In 2014 research was conducted in order to determine the driving cost of refuse collection.  It was determined that the main contributing factors were vehicle use and wages.  It was also determined that an increase in the amount of refuse collected often results in a reduction in operating cost.  It was determined that an increase of waste collection of 1% resulted in a decrease in cost of 0.2% per inhabitant.  In other words, the faster refuse can be collected on any given shift will reduce the costs associated with refuse collection.  This study highlights that efficiency is a critical factor in cost management (Greco, G., Allegrini, M., Del Lungo, C., Gori Savellini, P., & Gabellini, L., 2014).

            Greco, G. et al. (2014) also addressed an interesting finding in their study.  It was determined that the collection of undifferentiated waste has a higher cost advantage over that of differentiated waste.  Therefore, an increase in waste production has an economic incentive to collect only undifferentiated waste.  In order to limit this incentive, the waste collection process needs to be more efficient therefore limiting the incentive to collect only undifferentiated waste.

Recommended Solution

            In order to correct the issues we have today with our waste management and collection we must address two issues, hazards and efficiency.  Hazards are causing a large amount of reported on the job injuries costing our waste management systems both time and money.  The efficiency of our collection routes are not optimizing the full capabilities of our waste collection vehicles which again results in lost time and money.  We must immediately address these issues because as our population grows, they will only get worse.

            First we can address the hazard issues.  By eliminating the waste collection worker from the vehicle, we will practically eliminate the reports of on the job injuries.  We can do this by modifying our current waste collection vehicles with equipment and software that has been developed for autonomous transport vehicles.  In an article by Andreasson, H., Bouguerra, A., Cirillo, M., Dimitrov, D. N., Driankov, D., Karlsson, L., Stoyanov, T. (2015), they researched a system known as Safe Autonomous Navigation (SAUNA).  SAUNA has the capability to perceive its environment, allocate tasks with other vehicle in the system, plan motion and coordination, as well as provide predictive collision avoidance.  This system could create a completely autonomous environment for all collection vehicles within a system that would eliminate the need for workers to be exposed to the hazards of the waste collection process and potentially reduce injury reports by as much as 38% as reported by Kuijer, P. et al. (2010).

            Further studies have been done to address the mathematical computation required for such an autonomous system of vehicles.  In a study by Xidias, E., & Azariadis, P. (2011), it was concluded that multiple vehicles can work together in solving the same unique problem while at the same time avoiding collision with obstacles and each other during the process.  Martínez-Barberá, H., & Herrero-Pérez, D. (2010) conducted a similar study and stated that the development of a flexible and easily configurable and commanded autonomous ground vehicle could be easily adapted into any management system.

            Next we must address the issues of efficiency.  There are several studies that have been conducted that address topics such as demand responsive operations (Oyatoye, E., & J A O Magbagbeola, 2010), periodic vehicle routing (Campbell, A., & Wilson, J. 2014), inventory routing (Schutten, M., Pérez Rivera, A., & Mes, M., 2014), and simply re-evaluating vehicle route scheduling (Huang, S., & Lin, P., 2015).

      In the past, solid waste collection was carried out without analyzing demand and the construction of the routes was left to the drivers. Cities, however, continue to expand. Because of this ongoing urbanization, the importance of an efficient collection system only increases. Optimally, there should be a method that tries to maximize the general acceptance of a solution. However, as this is hard to realize, different methods have been developed that focus on route length, costs, number of collection vehicles, etc. (Belien, J., De, L., & Van, J. (2014).

            Another efficiency issue is the number of vehicles that are used for collection.  In the current model of manned vehicles, it is required to have more vehicles that necessary to collect on any given day in order to eliminate overtime pay of employees (Belien, J. et al., 2014).  By eliminating the waste collection worker from the equation we can in turn reduce the amount of collection vehicles required since these vehicle can collect essentially twenty-four hours a day.

            In a study by Benjamin, A., & Beasley, J. (2010), research was conducted that mathematically attempted to minimize the number of vehicles required for waste collection.  It is realized that having the ability for a vehicle to collect until full and then another vehicle be dispatched to continue collection could potentially reduce the number of waste collection vehicles required.  The issue with the idea is that in a manned vehicle scenario, it is nearly impossible to manage worker schedules.  Therefore, by implementing the unmanned waste collection vehicle, both a reduction in vehicles and more efficient collection could be maximized.

Conclusion

     Waste management is one of the major issues of urban engineering.  Today, the total amount of waste generated annually worldwide (municipal, industrial, hazardous) is more than 4 billion tons. Almost 45% of these are considered as municipal solid waste, while the rest is industrial waste, including hazardous one (Nakou, D., Benardos, A., & Kaliampakos, D. 2014).

            Further analysis must also be conducted into the future of selective waste collection (recycling).  By reducing the amount of waste being sent to landfills and in turn supplying the recycling industry, we can avoid the further degradation of our environment.  Other countries besides the U.S. have legal provisions that require residents to separate waste prior to disposal (da Silva Carvalho, M., Rosa, L., Bufoni, A., & de Sousa Ferreira, Araceli Cristina., 2011).  The U.S. must also look into the feasibility of adopting the same types of laws in order to improve our environment futures sake.

            The reductionist approach is unsustainable for the future of waste collection.  We must therefore adopt a more sustainable systems approach with the inclusion of Automated Unmanned Refuse Collection (AURC).  A sustainable waste management systems is one that is not dependent on expansion, focusses of the processes and not the products, organizational structures need to be adaptable and multipurpose, linking of transportation works to a mutual advantage, and feedback that effects system change (Seadon, 2010).  With the use of AURC, our waste management solutions will take us safely and economically into the future.

References

Andreasson, H., Bouguerra, A., Cirillo, M., Dimitrov, D. N., Driankov, D., Karlsson, L., Stoyanov, T. (2015). Autonomous transport vehicles: Where we are and what is missing. IEEE Robotics & Automation Magazine, 22(1), 64-75. doi:10.1109/MRA.2014.2381357

Belien, J., De, L., & Van, J. (2014). Municipal solid waste collection and management problems: A literature review. Transportation Science, 48(1), 78-102. doi:10.1287/trsc.1120.0448

Benjamin, A. M., & Beasley, J. E. (2010). Metaheuristics for the waste collection vehicle routing problem with time windows, driver rest period and multiple disposal facilities. Computers and Operations Research, 37(12), 2270-2280. doi:10.1016/j.cor.2010.03.019

Campbell, A. M., & Wilson, J. H. (2014). Forty years of periodic vehicle routing. Networks, 63(1), 2-15. doi:10.1002/net.21527

da Silva Carvalho, M., Rosa, L. P., Bufoni, A. L., & de Sousa Ferreira, Araceli Cristina. (2011). The issue of sustainability and disclosure. A case study of selective garbage collection by the urban cleaning service of the city of rio de janeiro, brazil – COMLURB. Resources, Conservation & Recycling, 55(11), 1030-1038. doi:10.1016/j.resconrec.2011.05.015

Greco, G., Allegrini, M., Del Lungo, C., Gori Savellini, P., & Gabellini, L. (2014). Drivers of solid waste collection costs. empirical evidence from italy. Journal of Cleaner Production, doi:10.1016/j.jclepro.2014.07.011

Huang, S., & Lin, P. (2015). Vehicle routing–scheduling for municipal waste collection system under the “Keep trash off the ground” policy. Omega, 55, 24-37. doi:10.1016/j.omega.2015.02.004

Kuijer, P. P. F. M., Frings-sen, M. H. W., & Sluiter, J. K. (2010). Health and safety in waste collection: Towards evidence-based worker health surveillance. American Journal of Industrial Medicine, 53(10), 1040-1064.

Martínez-Barberá, H., & Herrero-Pérez, D. (2010). Autonomous navigation of an automated guided vehicle in industrial environments. Robotics and Computer Integrated Manufacturing, 26(4), 296-311. doi:10.1016/j.rcim.2009.10.003

Oyatoye, E. O., & J A O Magbagbeola. (2010). A transportation model for demand responsive fleet operation in A manufacturing firm. The International Business & Economics Research Journal, 9(8), 29.

Schutten, M., Pérez Rivera, A., & Mes, M. (2014). Inventory routing for dynamic waste collection. Waste Management, 34(9), 1564-1576. doi:10.1016/j.wasman.2014.05.011

Seadon, J. K. (2010). Sustainable waste management systems. Journal of Cleaner Production, 18(16), 1639-1651. doi:10.1016/j.jclepro.2010.07.009

Tibbetts, J. (2013). Garbage collection is "one of the most hazardous jobs". CMAJ: Canadian Medical Association Journal, 185(7), 554. doi:10.1503/cmaj.109-4443

Nakou, D., Benardos, A., & Kaliampakos, D. (2014). Assessing the financial and environmental performance of underground automated vacuum waste collection systems. Tunnelling and Underground Space Technology, 41, 263. doi:10.1016/j.tust.2013.12.005

Xidias, E. K., & Azariadis, P. N. (2011). Mission design for a group of autonomous guided vehicles. Robotics and Autonomous Systems, 59(1), 34-43. doi:10.1016/j.robot.2010.10.003