Social and Economic Implications of Transport Sharing and Automation
Lead Research Organisation:
University of Glasgow
Department Name: School of Social & Political Sciences
Abstract
This study will link the changing nature of jobs due to automation and the platform economy to regional infrastructure planning and transport operations, and the role specifically of transport automation within this context. The patterns and forms of jobs are changing due to many different reasons, leading to non-traditional work schedules and differences in commuting patterns, non-standard work travel patterns, and even elimination of certain jobs and creation of new ones, with significant implications for regional infrastructure planning and transport operations. At the same time, there are enormous changes anticipated in infrastructure and operations, due to large-scale automation in the transport sector (eg autonomous and connected vehicles).
This project will make estimates of the changing nature of jobs due to these considerations at the regional level towards the goal of deriving the transport and regional infrastructural planning consequences. The project will use labour market survey data as well as privately-held labour market data on jobs, skills and industry to estimate regional variations due to these trends, given regional industry-occupation mix. These changes will be linked to the Spatial Urban Data System (SUDS), which is a UK-wide geospatial data infrastructure under development within UBDC containing transport infrastructural and operational conditions. , and which has been recently used to identify areas of transport poverty throughout the UK and the extent to which and which we will expand through work with the project's industrial partners.
Using these data sources, we will identify regional automation risks due to unique industry and skill concentrations and derive transport and infrastructure planning implications. Within this context, we will also evaluate the role of autonomous vehicles given potentially different commuting patterns using specialist transport simulation models. We will further develop specialist transport simulation models to ascertain which packages of "last-mile" transport solutions (low-energy station cars, autonomous vehicles, shared transport, active travel and demand-response services) are likely to bring about high-quality, sustainable and socially-equitable forms of transport accessibility in areas at risk of changing nature of jobs. We will then combine the results of our various model scenarios, using ensemble forecasting methods utilising Bayesian Model Averaging or related techniques to ascertain which packages are more likely to bring about high-quality transport accessibility in the selected areas.
This project will make estimates of the changing nature of jobs due to these considerations at the regional level towards the goal of deriving the transport and regional infrastructural planning consequences. The project will use labour market survey data as well as privately-held labour market data on jobs, skills and industry to estimate regional variations due to these trends, given regional industry-occupation mix. These changes will be linked to the Spatial Urban Data System (SUDS), which is a UK-wide geospatial data infrastructure under development within UBDC containing transport infrastructural and operational conditions. , and which has been recently used to identify areas of transport poverty throughout the UK and the extent to which and which we will expand through work with the project's industrial partners.
Using these data sources, we will identify regional automation risks due to unique industry and skill concentrations and derive transport and infrastructure planning implications. Within this context, we will also evaluate the role of autonomous vehicles given potentially different commuting patterns using specialist transport simulation models. We will further develop specialist transport simulation models to ascertain which packages of "last-mile" transport solutions (low-energy station cars, autonomous vehicles, shared transport, active travel and demand-response services) are likely to bring about high-quality, sustainable and socially-equitable forms of transport accessibility in areas at risk of changing nature of jobs. We will then combine the results of our various model scenarios, using ensemble forecasting methods utilising Bayesian Model Averaging or related techniques to ascertain which packages are more likely to bring about high-quality transport accessibility in the selected areas.
Planned Impact
With 66% of the world's population estimated to be living in urban areas by 2050, the need to provide new transport infrastructure and to address traffic congestion, road fatalities, air pollution, and associated problems continue to generate debates in policy circles.
Impact on local economic development and labour market planning: The work related to estimating the potential impact of automation and AI on jobs given skills required in the occupation-industry mix regionally available is likely to be of immense value to regional infrastructure planners and business owners, as well as for skills-development initiatives and organisations in the local economic development planning. Through UBDC's networks, we will engage local authorities and other stakeholders in co-creating our results on this topic, so as to involve our work in their planning processes.
Transport and infrastructure planning and operations impact: Additionally, there is a range of shared technology, automation and use of AI and Machine Learning (ML) being proposed in transport, and a growing business community involved in their development and use. Although the trends surrounding automation and sharing transport are being driven primarily by private companies, governments around the world are increasingly developing policies to support as well as to regulate many of these developments, and the UK government has an active programme on Connected and Autonomous Vehicles (CAV); among the high-value economic infrastructure to be funded through the 2016 National Productivity Investment Fund (NPIF) is £390 million for future transport including ultra-low emission vehicles and CAVs. Additionally, various types of shared mobility services, e.g., car-sharing, dynamic ride-sharing, on-demand personal mobility vans, and express, crowd-sourced urban delivery services, under the banner of Mobility-As- A-Service (MaaS) are now offered by private companies in UK cities. Just as the introduction of the private car in the beginning of the twentieth century transformed the way we live our daily lives and the ways in which cities developed, large-scale automation, connectivity and sharing of mobility resources (sharing economy) are expected to be a step-change changing daily lives in future society and the form and functions in cities.We expect that our results will help highlight significant regional planning and operations impact of such technology, against the backdrop of changing commuting and work-related travel patterns resulting from the changing nature of jobs.
Industrial Impacts: The approach will allow us to evaluate spatial and regional effects of varying degrees of automation and sharing mobility, and to identify new markets in smart cities and urban planning. Our industry partner, Peter Brett Associates, views that the risks of emerging technology being left out of the planning agenda are great due to lack of empirical data, leading technology disruption in transport to occur in an ad-hoc way. Having the results of the analysis and the associated data would help them reach new markets and also to reduce uncertainty in their forecasts. Scottish MaaS, has similarly noted that the methods and results being proposed will help them reach new markets both geographically thereby opening up UK companies to a global pipeline of contracts in integrating CAVs into infrastructure planning and construction, and MaaS solutions in addressing expensive last-mile problems facing city managers worldwide.
Impact on local economic development and labour market planning: The work related to estimating the potential impact of automation and AI on jobs given skills required in the occupation-industry mix regionally available is likely to be of immense value to regional infrastructure planners and business owners, as well as for skills-development initiatives and organisations in the local economic development planning. Through UBDC's networks, we will engage local authorities and other stakeholders in co-creating our results on this topic, so as to involve our work in their planning processes.
Transport and infrastructure planning and operations impact: Additionally, there is a range of shared technology, automation and use of AI and Machine Learning (ML) being proposed in transport, and a growing business community involved in their development and use. Although the trends surrounding automation and sharing transport are being driven primarily by private companies, governments around the world are increasingly developing policies to support as well as to regulate many of these developments, and the UK government has an active programme on Connected and Autonomous Vehicles (CAV); among the high-value economic infrastructure to be funded through the 2016 National Productivity Investment Fund (NPIF) is £390 million for future transport including ultra-low emission vehicles and CAVs. Additionally, various types of shared mobility services, e.g., car-sharing, dynamic ride-sharing, on-demand personal mobility vans, and express, crowd-sourced urban delivery services, under the banner of Mobility-As- A-Service (MaaS) are now offered by private companies in UK cities. Just as the introduction of the private car in the beginning of the twentieth century transformed the way we live our daily lives and the ways in which cities developed, large-scale automation, connectivity and sharing of mobility resources (sharing economy) are expected to be a step-change changing daily lives in future society and the form and functions in cities.We expect that our results will help highlight significant regional planning and operations impact of such technology, against the backdrop of changing commuting and work-related travel patterns resulting from the changing nature of jobs.
Industrial Impacts: The approach will allow us to evaluate spatial and regional effects of varying degrees of automation and sharing mobility, and to identify new markets in smart cities and urban planning. Our industry partner, Peter Brett Associates, views that the risks of emerging technology being left out of the planning agenda are great due to lack of empirical data, leading technology disruption in transport to occur in an ad-hoc way. Having the results of the analysis and the associated data would help them reach new markets and also to reduce uncertainty in their forecasts. Scottish MaaS, has similarly noted that the methods and results being proposed will help them reach new markets both geographically thereby opening up UK companies to a global pipeline of contracts in integrating CAVs into infrastructure planning and construction, and MaaS solutions in addressing expensive last-mile problems facing city managers worldwide.
People |
ORCID iD |
| Long Chen (Principal Investigator / Fellow) |
| Description | We developed a multi-level Wavenet framework that take a series of meta-data features into account so as to effectively capture the varying demand patterns. Specifically, the model is a one-dimensional convolutional neural network that includes two sub-networks that aims to encode the source series and decode the predicting series, respectively. The two sub-networks are combined by stacking the decoder on top of the encoder, which in turn, preserves the temporal patterns of the time series. Experiments on large-scale real taxi demand dataset of NYC demonstrate that our model is highly competitive to the existing ones. |
| Exploitation Route | Traffic demand prediction is at the core of intelligent transportation systems when developing a smart city. Accurately predicting taxi demand can help the city manager/operater to optimize the resources, and thus reducing the drivers' waiting/idle time which waste energy and cause traffic congestion. However, exploiting traffic time series to facilitate the demand prediction is a thorny problem since traffic demand usually unevenly distributed over time and space. |
| Sectors | Digital/Communication/Information Technologies (including Software),Transport |
| Description | IntelligentTraffic Modelling in NYC |
| Organisation | Rutgers University |
| Department | Rutgers Urban and Civic Informatics Lab |
| Country | United States |
| Sector | Academic/University |
| PI Contribution | The biggest contribution that I made was creating a new intelligent traffic system, introducing UBERNET - a deep learning based predictive system, and then help them to set it up on the local NYC data. |
| Collaborator Contribution | Their data scientists are mostly responsible for collecting and cleaning the traffic and social media dataset, conducting some preliminary analysis about the explanatory features, and carrying out a data engineering job to convert it into the format that is recognizable to our program. |
| Impact | A new dataset is created for predicting the taxi pickups in NYC, which combines a series of factors from heterogeneous sources. This is arguably the first dataset that covers four varying feature set in NYC for the prediction: [A] temporal and real-time features (e.g. time-of-day, day-of-week, and hourly weather conditions from Meteoblue (https://www.meteoblue.com), [B] averages of demographic and socioeconomic features such as income levels and unemployment of census tracks where pick-ups occurred during the time interval, from the American Community Survey 2008-2011 5-year sample (https://www.census.gov/programs-surveys/acs) [C] average travel-to-work characteristics such as the proportion of commuters who walk or take public transit in the census tracks in which the pickups are located during the interval, also from the census data and [D] social and built environment characteristics of the census tract where the pickup occurred during the period, for example of crime levels from the NYC police department (https://www1.nyc.gov/site/nypd/stats/crime-statistics/historical.page), and data on the built environment (e.g. density of transport facilities such as the number of transport stops, stations and other facilities) from the NYC Planning Department \footnote{\url{https://www1.nyc.gov/site/planning/data-maps/open-data/dwn-selfac.page}}.} |
| Start Year | 2020 |
| Description | An Evening with Uber |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Professional Practitioners |
| Results and Impact | 200 people attended this Evening Event with Uber, where we discussed the possibility of predicting Uber demand at local and global level, thereby improving the design of the intelligent traffic system. |
| Year(s) Of Engagement Activity | 2020 |
| Description | Traffic Scotland Event |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Professional Practitioners |
| Results and Impact | Around 80 people attended this event, where I gave a presentation about taxi flow prediction using deep learning based approaches, which could further optimize the current intelligent system. Many audiences are intrigued later on and asked for more detailed information. |
| Year(s) Of Engagement Activity | 2020 |
| Description | Transportation Research Board |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Industry/Business |
| Results and Impact | I made an oral presentation at the international conference Transportation Research Board 2020, where people from industry such as Uber and Toyota asked questions with respect to the feasibility of applying our traffic prediction approach on their dataset, many people from traffic community also interested in our approach with constructive suggestions. |
| Year(s) Of Engagement Activity | 2020 |
| Description | U21 ERC Workshop |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Schools |
| Results and Impact | 150 researchers in the transport community attended the workshop, which sparked questions and discussion afterwards, with increased from several cohorts who are interested in using deep learning approach to facilitate transportation system. |
| Year(s) Of Engagement Activity | 2018 |