Data analytics for smart parking applications

22Citations
Citations of this article
87Readers
Mendeley users who have this article in their library.

Abstract

We consider real-life smart parking systems where parking lot occupancy data are collected from field sensor devices and sent to backend servers for further processing and usage for applications. Our objective is to make these data useful to end users, such as parking managers, and, ultimately, to citizens. To this end, we concoct and validate an automated classification algorithm having two objectives: (1) outlier detection: to detect sensors with anomalous behavioral patterns, i.e., outliers; and (2) clustering: to group the parking sensors exhibiting similar patterns into distinct clusters. We first analyze the statistics of real parking data, obtaining suitable simulation models for parking traces. We then consider a simple classification algorithm based on the empirical complementary distribution function of occupancy times and show its limitations. Hence, we design a more sophisticated algorithm exploiting unsupervised learning techniques (self-organizing maps). These are tuned following a supervised approach using our trace generator and are compared against other clustering schemes, namely expectation maximization, k-means clustering and DBSCAN, considering six months of data from a real sensor deployment. Our approach is found to be superior in terms of classification accuracy, while also being capable of identifying all of the outliers in the dataset.

Figures

  • Figure 1. Probability distribution function of the parking duration (random variable (r.v.) tON).
  • Table 1. Average, Max and Min values of the Weibull functions we fit on the dataset.
  • Figure 2. Holiday/weekend vs. weekday occupancy rates in the month of January 2015: comparison between occupancy rates from real data and synthetic traces generated by our event-based simulator. (a) Worldsensing dataset; (b) event-based simulator.
  • Figure 3. Anomaly detection: accuracy vs. detection rate: month of January 2015. Different curves correspond to specific increases in the parking times, from 10% to 70% for 20% of the parking nodes.
  • Figure 4. Feature function f (t): calculation example for a typical parking sensor. The feature vector xi = [xi1, . . . , xiK ]T for sensor i is obtained as xit ← f (t) for t = 1, . . . , K.
  • Figure 5. Flow diagram of the SOM training process.
  • Figure 6. Flow diagram of the SOM training process.
  • Figure 7. Weighted F-measure for the four selected clustering techniques vs. the number of clusters k.

References Powered by Scopus

Least Squares Quantization in PCM

11513Citations
N/AReaders
Get full text

Data clustering: A review

10810Citations
N/AReaders
Get full text

Data clustering: 50 years beyond K-means

7286Citations
N/AReaders
Get full text

Cited by Powered by Scopus

Smart parking in IoT-enabled cities: A survey

260Citations
N/AReaders
Get full text

Big Data for Internet of Things: A Survey

251Citations
N/AReaders
Get full text

Analytics for the internet of things: A survey

162Citations
N/AReaders
Get full text

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Cite

CITATION STYLE

APA

Piovesan, N., Turi, L., Toigo, E., Martinez, B., & Rossi, M. (2016). Data analytics for smart parking applications. Sensors (Switzerland), 16(10). https://doi.org/10.3390/s16101575

Readers over time

‘16‘17‘18‘19‘20‘21‘22‘23‘240481216

Readers' Seniority

Tooltip

PhD / Post grad / Masters / Doc 33

79%

Researcher 6

14%

Lecturer / Post doc 2

5%

Professor / Associate Prof. 1

2%

Readers' Discipline

Tooltip

Computer Science 25

49%

Engineering 24

47%

Social Sciences 1

2%

Medicine and Dentistry 1

2%

Article Metrics

Tooltip
Social Media
Shares, Likes & Comments: 15

Save time finding and organizing research with Mendeley

Sign up for free
0