A multi-modal social force model for conflict analysis and assessment of traffic safety and capability of shared spaces

该项目的目的是为所有交通模式开发运动和预测冲突避免行为模型，以便真实地复制共享空间中道路使用者的复杂行为。这允许开发一种工具，可以在规划过程的设计阶段早期协助估计这类交通区域的交通安全。具体考虑到冲突解决策略方面，在第一个资助期，开发了一个多层模型，该模型包括对人类感知的现实近似，并为每种冲突情况选择特定的冲突解决策略。利用该模型进行的模拟得出了可行的结果，这些结果已经通过使用宏观交通参数对真实世界数据进行了验证。然而，对交通状况的进一步分析表明，特别是骑自行车的人在许多情况下表现出广泛的可能的行为方式。考虑到尽可能多的不同的行为方式对于交通安全的检查是至关重要的。当前的确定性模型不允许同时考虑多种行为方式。因此，在拟议的项目中，所有可观察到的冲突解决策略将从现实世界的数据中提取出来，并确定和分析导致选择特定策略的因素。基于这些因素，建立了冲突解决策略选择的随机决策模型，并与策略相结合，整合到整体模型中。此外，决策模型考虑了人类感知中的不确定性。在第一个供资期间的研究表明，不仅在冲突情况下的行为，而且自由流动也是不同外部因素的基础。由于上一个筹资期间的重点是分析冲突局势，因此对这些影响进行了定性检测，但没有详细地加以参数化。但是，为了全面评估交通安全，必须对所有正在发生的情况，包括自由流动和冲突情况进行评估。为了满足仿真模型的这些要求，在建议的项目中，确定、量化控制自由流运动的因素，然后用于自动生成真实的自由流轨迹。最后，开发安全测量方法来估计共享空间交叉口布局的交通安全性。结合扩展模型，可以根据仿真结果对特定交通区域的安全性进行评估。

The aim of this project is the development of motion and predictive conflict avoidance behavior models for all traffic modes, in order to realistically replicate the complex behavior of road users in shared spaces. This allows for the development of a tool that can assist with the estimation of traffic safety in this type of traffic areas as early as the design stage of a planning process.Specifically considering the aspect of conflict solving strategies, in the first funding period, a multi-layer model was developed that comprises a realistic approximation of the human perception and results in the choice of a specific conflict solving strategy for each conflict situation. Simulations utilizing this model lead to feasible results that have been validated against real-world data using macroscopic traffic parameters. However, further analysis of traffic situations revealed that especially cyclists present a broad spectrum of possible ways of behavior in many situations. The consideration of as many different ways of behavior as possible is of paramount importance for an examination of traffic safety. The current deterministic model does not allow for the simultaneous consideration of multiple ways of behavior. For this reason, in the proposed project all observable conflict solving strategies will be extracted from the real-world data and the factors which lead to the choice of a specific strategy are identified and analyzed. Based on these factors, a stochastic decisional model for the choice of conflict solving strategies is developed and, in connection with the strategies, integrated into the overall model. Additionally, the decisional model considers uncertainties in human perception.Research during the first funding period revealed that not only the behavior in a conflict situation but also free flow movements underlie different external factors. Since in the last funding period the focus was on the analysis of conflict situations, these effects were qualitatively detected, but not parametrized in detail. However, for a comprehensive assessment of traffic safety an evaluation of all occurring situations, both free flow movements as well as conflict situations, is necessary. To meet these requirements in respect to a simulation model, in the proposed project, the factors that govern free flow movements are determined, quantified and then used to automatically generate realistic free flow trajectoriesFinally, safety measurements are developed for estimating the traffic safety of shared space intersection layouts. In conjunction with the extended model, it is possible to assess the safety of specific traffic areas based on simulation results.