Simulation systems have become an essential component in the development and validation of autonomous driving technologies. The prevailing state-of-the-art approach for simulation is to use game engines or high-fidelity computer graphics (CG) models to create driving scenarios. However, creating CG models and vehicle movements (a.k.a. the assets for simulation) remains a manual task that can be costly and time-consuming. In addition, the fidelity of CG images still lacks the richness and authenticity of real-world images and using these CG images for training leads to degraded performance.In this paper we present a novel approach to address these issues: Augmented Autonomous Driving Simulation (AADS). Our formulation augments real-world pictures with a simulated traffic flow to create photo-realistic simulation images and renderings. More specifically, we use LiDAR and cameras to scan street scenes. From the acquired trajectory data, we generate highly plausible traffic flows for cars and pedestrians and compose them into the background. The composite images can be re-synthesized with different viewpoints and sensor models (camera or LiDAR). The resulting images are photo-realistic, fully annotated, and ready for end-to-end training and testing of autonomous driving systems from perception to planning. We explain our system design and validate our algorithms with a number of autonomous driving tasks from detection to segmentation and predictions.Compared to traditional approaches, our method offers unmatched scalability and realism. Scalability is particularly important for AD simulation and we believe the complexity and diversity of the real world cannot be realistically captured in a virtual environment. Our augmented approach combines the flexibility of a virtual environment (e.g., vehicle movements) with the richness of the real world to allow effective simulation of any location on earth. Summary

Figure 1: Our heterogeneous multi-agent simulation algorithm can be used for scenarios with tens or hundreds of different types of agents sharing a physical space. Pedestrians walking on a street (the first). Cars moving on a twisting road (the second). Traffic including cars and pedestrians (the third). Traffic shown through VR (the fourth). Our approach can generate plausible behaviors at interactive rates on a desktop PC and through VR. ABSTRACTInteractive multi-agent simulation algorithms are used to compute the trajectories and behaviors of different entities in virtual reality scenarios. However, current methods involve considerable parameter tweaking to generate plausible behaviors. We introduce a novel approach (Heter-Sim) that combines physics-based simulation methods with data-driven techniques using an optimization-based formulation. Our approach is general and can simulate heterogeneous agents corresponding to human crowds, traffic, vehicles, or combinations of different agents with varying dynamics. We estimate motion states from real-world datasets that include information about position, velocity, and control direction. Our optimization algorithm considers several constraints, including velocity continuity, collision avoidance, attraction, and direction control. To accelerate the computations, we reduce the search space for both collision avoidance and optimal solution computation. Heter-Sim can simulate tens or hundreds of agents at interactive rates and we compare its accuracy with real-world datasets and prior algorithms. We also perform user studies that evaluate the plausible behaviors generated by our algorithm and a user study that evaluates the plausibility of our algorithm via VR.

We present a novel data-driven approach to populate virtual road networks with realistic traffic flows. Specifically, given a limited set of vehicle trajectories as the input samples, our approach first synthesizes a large set of vehicle trajectories. By taking the spatio-temporal information of traffic flows as a 2D texture, the generation of new traffic flows can be formulated as a texture synthesis process, which is solved by minimizing a newly developed traffic texture energy. The synthesized output captures the spatio-temporal dynamics of the input traffic flows, and the vehicle interactions in it strictly follow traffic rules. After that, we position the synthesized vehicle trajectory data to virtual road networks using a cage-based registration scheme, where a few traffic-specific constraints are enforced to maintain each vehicle's original spatial location and synchronize its motion in concert with its neighboring vehicles. Our approach is intuitive to control and scalable to the complexity of virtual road networks. We validated our approach through many experiments and paired comparison user studies.

Improving the urban water allocation system has depended on the different technical and managerial components. One of the main solutions for increasing the available capacity is the predetermined plan for constructing a water circulation system to restore the existing water resources. For this objective, a combination model was conducted based on the storm-water management technique (SMT) and environment landscape system (ELS) under drought condition. A hydrological framework was established using the long-term meteorological information in central China to estimate the extreme values of surface water in each stress period. Data analysis system was generated at three meteorological point of the study area and the developed model was incorporated to simulate the behavior of the subsurface flow. Consequently, a growth simulation model was designed according to the soil structure and vegetation canopy cover to formulate a real-time plan for the restoration of ELS. Results showed that the proposed model could be improved the water productivity in urban and environment consumptions. Furthermore, the technical analysis confirmed the suitability and applicability of the developed plan in cities with water shortage.

We present a biologically plausible dynamics model to simulate swarms of flying insects. Our formulation, which is based on biological conclusions and experimental observations, is designed to simulate large insect swarms of varying densities. We use a force-based model that captures different interactions between the insects and the environment and computes collision-free trajectories for each individual insect. Furthermore, we model the noise as a constructive force at the collective level and present a technique to generate noise-induced insect movements in a large swarm that are similar to those observed in real-world trajectories. We use a data-driven formulation that is based on pre-recorded insect trajectories. We also present a novel evaluation metric and a statistical validation approach that takes into account various characteristics of insect motions. In practice, the combination of Curl noise function with our dynamics model is used to generate realistic swarm simulations and emergent behaviors. We highlight its performance for simulating large flying swarms of midges, fruit fly, locusts and moths and demonstrate many collective behaviors, including aggregation, migration, phase transition, and escape responses.