Academia.eduAcademia.edu

Figure 5 – uploaded by Andreas Keler

See full PDFdownloadDownload figure
Figure 5. Applied desired acceleration function.

Figure 5 Applied desired acceleration function.

Related Figures (9)

Table 1. Parameter values from the COWI report. Trajectories extracted from areal video data are limited in length and typically permit the analysis of behavior over a relatively short road segment or intersection. The detailed trajectory output from simulator studies makes it possible to analyze the interaction behavior of one road user over a long time period. Such data have been used previously to calibrate following models for car drivers [20].
Table 1. Parameter values from the COWI report. Trajectories extracted from areal video data are limited in length and typically permit the analysis of behavior over a relatively short road segment or intersection. The detailed trajectory output from simulator studies makes it possible to analyze the interaction behavior of one road user over a long time period. Such data have been used previously to calibrate following models for car drivers [20].
Figure 1. Vectors to identify following situations. (a) First cosine similarity measure. (b) Second cosine similarity measure. The vectors and angles used to derive following situations are shown in Figure 1a,b. Using the two cosine similarities defined in Equation (1) and Equation (4) and the distance between the road users Equation (2), the following intervals Tj; = =(t ij_by tij | for each road user pair ij are determined. The beginning of the following ‘ataivall tij_p is de- fined to be the first moment in which djj(t) < 30 m and cos @(t) > 0.75 and cos a(t) > 0.95. The endpoint of the following interval t;; . is defined to be when dj;(t) > 1.1 x 30m or cos 6(t) < 0.0. Multiple following intervals can exist for each pair ij.
Figure 1. Vectors to identify following situations. (a) First cosine similarity measure. (b) Second cosine similarity measure. The vectors and angles used to derive following situations are shown in Figure 1a,b. Using the two cosine similarities defined in Equation (1) and Equation (4) and the distance between the road users Equation (2), the following intervals Tj; = =(t ij_by tij | for each road user pair ij are determined. The beginning of the following ‘ataivall tij_p is de- fined to be the first moment in which djj(t) < 30 m and cos @(t) > 0.75 and cos a(t) > 0.95. The endpoint of the following interval t;; . is defined to be when dj;(t) > 1.1 x 30m or cos 6(t) < 0.0. Multiple following intervals can exist for each pair ij.
Figure 2. Examples of observed passing and following situations. Examples of the three types of following /passing behavior observed in the simulator study are shown in Figure 2.
Figure 2. Examples of observed passing and following situations. Examples of the three types of following /passing behavior observed in the simulator study are shown in Figure 2.
Figure 3. Distribution for the headway time parameter CC1. The headway time CC1 is estimated for each observed following /passing events by defining Safety distance = AX,;, during the following phase (blue in Figure 2) and the speed of the simulator study participant at that point. CCO was set at 0.8 m based on the findings of [6]. The resulting cumulative density function of CC1 is shown in Figure 3.
Figure 3. Distribution for the headway time parameter CC1. The headway time CC1 is estimated for each observed following /passing events by defining Safety distance = AX,;, during the following phase (blue in Figure 2) and the speed of the simulator study participant at that point. CCO was set at 0.8 m based on the findings of [6]. The resulting cumulative density function of CC1 is shown in Figure 3.
Table 2. Descriptive statistics of the parameter estimates. 4.3. Validation
Table 2. Descriptive statistics of the parameter estimates. 4.3. Validation
Figure 4. Observed points for following/ passing events (left) and following/exiting events (right).
Figure 4. Observed points for following/ passing events (left) and following/exiting events (right).
Table 3. Parameters sets.
Table 3. Parameters sets.
Table 4. Evaluation of parameter sets based on tp picycie: where tp is the total queue dissipation time and n is the number of bicyclists in the queue. The percent error and results of a t-test to determine significant differences in the means between the simulated and observed data are listed in Table 4. Bold text indicates that no significant difference between the means of the observed and simulated data was found.
Table 4. Evaluation of parameter sets based on tp picycie: where tp is the total queue dissipation time and n is the number of bicyclists in the queue. The percent error and results of a t-test to determine significant differences in the means between the simulated and observed data are listed in Table 4. Bold text indicates that no significant difference between the means of the observed and simulated data was found.
Figure 6. Observed and simulated average queue dissipation time. Boxplots of the average queue dissipation time from the simulation runs and the field observations are shown in Figure 6. The horizontal bar in the boxplots represents the median and the white rectangle indicates the mean. The capacities at the stop line delivered by the COWI report parameter set range be tween 1662 and 2268 bicyclists/h/m while those found using the trajectory data paramete! set lie between 924 and 1250 bicyclists/h/m. Interestingly the highest capacities per mete! width are found for the 1.8 m wide facility with both parameter sets.
Figure 6. Observed and simulated average queue dissipation time. Boxplots of the average queue dissipation time from the simulation runs and the field observations are shown in Figure 6. The horizontal bar in the boxplots represents the median and the white rectangle indicates the mean. The capacities at the stop line delivered by the COWI report parameter set range be tween 1662 and 2268 bicyclists/h/m while those found using the trajectory data paramete! set lie between 924 and 1250 bicyclists/h/m. Interestingly the highest capacities per mete! width are found for the 1.8 m wide facility with both parameter sets.
Related topics:
Transportation EngineeringIntelligent Transportation SystemsCar-Following ModelMicroscopic Traffic SimulationBicycle and Pedestrian TransportationBicycle TrafficMicroscopic SimulationTechnical University of MunichCar-Following ModelsBicycle Simulator
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved