Data Analysis
The accelerometer gives 5 distinct measures. The 5 measures include latitude, longitude, x, y, and z. The cell phone is placed vertically in such a way that the beginning coordinates begin at 0,10,0. X is the shifting between right and left, Y is the direction of motion of the vehicle, and Z is the direction which acts vertically upwards to the direction of motion. The values of these axes are given in m/s^2 . Runs were conducted from the same origin and destination to be able to identify similar patterns. Different approaches were taken to try and identify these patterns. For this study three runs are utilized. First step was making the run, pressing start once the ran began and ending it when the run was finalized. Next hitting transfer to be able to display the data on the website (http://smartcitycrowd.000webhostapp.com/index.php). Once the data was uploaded on the website, it was organized on an excel spreadsheet. The five original columns include the original measures, while other measures where derived from that data set. For example, x^2, ∆x, the standard deviation of x and so forth with the variables y and z. The tool used for analyzing and comparing these data points include MiniTab and ARCGis component called ARCMap. Minitab is used to plot the data points to be able to generate graphs and ARCMap is used to be able to visualize the points by using longitude and latitude on actual street map. A total of nine graphs were generated to be able to identify a pattern for these graphs. Below is a screenshot of the total distance of the runs once they are inserted onto ARCMap.
Figure 1: Geospatial Data Analysis
The white star represents the start of the run (my home) and the yellow start represents the end of the run (University of Texas at El Paso). Due to the overlapping of the three runs, they are not all visible though they are inserted onto ARCMap.
For the following visualization process, the time is measured in seconds. Since there are 5 readings per second, 1 second is shown in the 5th mark. For example, the graphs show 1000 total readings which represent a segment of 200 seconds. Data was analyzed once the I-10 segment was entered through McRae Blvd to be able to maintain a constant variable of speed. An average speed of 95 km per hour or 26.4 meters per second reaches an approximate total distance of 5300 (5277.8) meters. All graphs include a legend depicted as 1, 2, 3 each representing a run.
Figure 2: Starting segment of the geospatial data
The image above was screenshotted through ARCMap, we can see the beginning of the analyzed segment. Data points were searched using the ARCMap report, making sure both latitude and longitude coincided for the three runs. This process was used to have an accurate starting point. The dots represent a segment of 1 second and are color coded between each run. We can observe purple, green, and blue dots.
For the graph shown below we can observe the x values for 200 seconds. The peak points, using absolute values were taken from the three runs. Minitab gives us an option to select these points. Once these points were selected they were recorded to to beable to find the exact points on ARCMap.
Graph 1: X vs time
Using ArcMap the highest points were searched for on each individual run report, to see if they coincided in the same area. For this particular x graph which represents the shift between left and right happened to coincide at about the 1000 m mark from the starting point. Below is a screenshotted image of the coinciding points. Points also coincided at at 2650 m mark. The images below display the different distances.
Figure 3: 1000 meter mark (x values)
Figure 2: 2650 meter mark (x values)
The graph below shows the direction of motion, that is why we can see that it is much more stable through the 200 seconds. This variable begins at 10 for the x because of the phone’s vertical orientation. Once peak points were inserted into ARCMap points were not coinciding at a certain point. This could be due to sudden braking. We can see a cluster of high points at the beginning when average speed is barely being reached. Some of the images shown below are screenshotted to demonstrate some of the peak points for run 3 which seemed to have the highest peak points in the first 400 meters of the run.
Graph 2: Y vs Time
This image from ARCMap shows peak points at an approximate distance of 400 meters from the starting point.
Figure 4: 400 meter mark (y values)
The graph below shows the z values, which are the values that vary with the direction of motion of gravity therefore the up and down movements such as bumps, holes, or simple changes in elevation. From this graph we can see that though the starting point of this graph should be at 0, there is not one point beginning from there. We see similar pattern changes in this graph for all three runs. Peak points range from the 250 meter mark to the 1500 meter mark. As one drives through this intersection of the freeway, one can see the way there is a section with rough pavement such as changes in the pavement types.
Graph 3: Z vs Time
Figure 5: Pavement changes (z values)
The Image shown above shows the section of the freeway running on I-10 west which has a section of rough pavement. We also experience changes in elevation as there is a separation of pavement when making a lane change which could have been the cause of the peak points throughout this distance.
The following graphs are just used to demonstrate the emphasis of the peak points for the graphs shown below; All values are squared.
Graph 4: x^2 vs time
Graph 5: y^2 vs time
Graph 6: z^2 vs time
The following graphs are the standard deviation graph for eahc individual reading. The standard deviation was changing greatly throughout. This could have been experienced due to inaccuracy in the placement of the phone which could have increased the vibration throughout the runs.
Graph 7: Standard Deviation of x
Graph 8: Standard Deviation of y
Graph 9: Standard Deviation of z
In conclusion, the use of the accelerometer can be used to gather information such as changes in road conditions, behavior of driver (ex. sudden braking), and direction of driver can be found. For future works, one can analyze smaller segments repeating the same route to test the accuracy of accelerometer. Also try to analyze another measurement to see if further relationships can be found between all three variables.