AI Summary Reader Response Draft #2

Artificial intelligence (AI) and machine learning (ML) are incredible technologies with endless applications such as voice assistance, autonomous vehicles, e-commerce and many more. The article “How Artificial Intelligence, Robots Enhance Forest Sustainability in Finland“ by Scot McQueen (2019) discusses the use of these technologies to study forests in Finland, where data and information are collected from aerial and lidar imagery and then analysed using AI. Different indicators such as forest stands, species and inventories are predicted, providing better insights on the forests. McQueen reported that with insights done by AI, the Finish Forest Centre now has better awareness on how to take care of Finland’s forests while balancing the country's needs of using its forests for sustainable economic return. 

While McQueen discussed the use of of these technologies in forests, it is also used in other sectors such as finance, healthcare and engineering. A potential use for these technologies are in public transportation; particularly in traffic management. Traffic flow and congestions can be monitored and analysed using ML and the insights provided can help reduce congestions, carbon emissions, and ultimately improve the economy of the city. However, while AI and ML can be useful, there are limitations to these technologies.

ML can be used to improve roads and traffic flow. Advancements in wireless vehicle communication technology as well as GPS data presents new opportunities in identifying and predicting congestion patterns using various ML algorithms (Elfar et al., 2018). Analysed data can be applied to warn drivers ahead of traffic slowdowns, reducing potential accidents. In addition, Talebpour et al. (2013) stated that location of congestions can also be predicted, in which speed limits upstream can be applied.

Another application of ML is the reduction of carbon emissions. Globally, around one quarter of emissions come from the transportation sector, with two-thirds coming from road travel (IPCC, 2018). ML can assist in the reduction of unnecessary long trips on the road by optimizing vehicle routing and freight travel (Zeng et al., 2017). Lesser vehicles on the road equates to the reduction of carbon emissions. In addition, ML can influence vehicle engineering to become more efficient which also helps with emission levels. Ali (2018) revealed that hybrid vehicles can benefit from improved power management methods with the help of ML.

With lesser congestions, the productivity and economy for a city can improve. According to a 2017 report by INRIX, a company specializing in location-based data and analytics; over the next decade the most congested 25 cities of the U.S. are estimated to cost the drivers $480 billion due to lost time, wasted fuel, and carbon emitted due to traffic congestions (Pishue, 2017). This loss is also affecting the global economy to a great extent. Fortunately, data and insights from ML can be used to target congestions on the road, allowing for smoother traffic flow, lesser vehicles on the road and reduced carbon emissions. This will ultimately improve the economy.

However, the drawbacks of using ML come from inaccuracies of data collection. Stationary sensors such as highway cameras may face interruptions, reducing accuracy of spatiotemporal data of traffic. GPS data do not specify vehicle type and size, making it difficult to determine the distribution of vehicles on the road. Lastly, outliers such as pedestrians and cyclists may also affect the dataset (Aktar & Moridpour, 2021). Inaccurate data negatively impacts analysis done by ML, resulting in unhelpful insights.

Another drawback of ML algorithms are its limitations when predicting insights. Ali and Söffker (2018) reported that using the regression model machine learning algorithms are efficient in solving congestion forecasting problems, but are sensitive to outliers. Another algorithm, the support vector machine, is efficient in pattern recognition and classification, but takes high computational time and memory. It is important to choose the appropriate algorithm when analysing a specific road network. While an algorithm may be useful in one road network, it may not be as accurate when modelled to another.

In a nutshell, utilizing ML to reduce traffic congestions is feasible and will have a positive influence on society, the economy, and the environment. More sensors will need to be installed on road networks to ensure accurate data collection, and data scientists and engineers must refine ML techniques for improved congestion analyses. The government must also be involved so that road and township planning is optimized in reducing congestions.


[Updated on 18/2/22]

[Updated on 20/2/22]


References

Akhtar, M., & Moridpour, S. (2021). A review of traffic congestion prediction using artificial intelligence. Journal of Advanced Transportation, 2021.

Ali, A. M., & Söffker, D. (2018). Towards optimal power management of hybrid electric vehicles in real-time: A review on methods, challenges, and state-of-the-art solutions. Energies, 11(3), 476.

Elfar, A., Talebpour, A., & Mahmassani, H. S. (2018). Machine learning approach to short-term traffic congestion prediction in a connected environment. Transportation Research Record, 2672(45), 185-195.

Fleming, S. (2019, March 7). Traffic congestion cost the US economy nearly $87 billion in 2018. World Economic Forum
https://www.weforum.org/agenda/2019/03/traffic-congestion-cost-the-us-economy-nearly-87-billion-in-2018

IPCC. (2018). Global warming of 1.5°C. https://www.ipcc.ch/site/assets/uploads/sites/2/2019/06/SR15_Full_Report_High_Res.pdf

Pishue, B. (2017). US Traffic Hot Spots: Measuring the Impact of Congestion in the United States.

Talebpour, A., Mahmassani, H. S., & Hamdar, S. H. (2013). Speed harmonization: Evaluation of effectiveness under congested conditions. Transportation research record, 2391(1), 69-79.

Zeng, W., Miwa, T., & Morikawa, T. (2017). Application of the support vector machine and heuristic k-shortest path algorithm to determine the most eco-friendly path with a travel time constraint. Transportation Research Part D: Transport and Environment, 57, 458-473.

 

Comments

Popular Posts