Local Collision Research Articles

Local Collision Avoidance Algorithm for a Unmanned Surface Vehicle Based on Steering Maneuver Considering COLREGs

In order to sail safely in a marine environment with neighboring obstacles, a local collision avoidance algorithm based on a steering maneuver considering International Regulations for Preventing Collisions at Sea (COLREGs) is proposed for unmanned surface vehicles (USVs). In this paper, the algorithm consists of three parts: collision risk assessment, steering occasion determination, and navigation waypoint update. The collision risk assessment is used to judge a collision risk based on an analytical feasible angle interval by the closest point of approach method. The steering occasion determination provides a safe distance interval from an obstacle by analyzing a marginal steering angular velocity. The navigation waypoint update generates a temporary waypoint to navigate the USV and considers the overtaking, head-on, and crossing rules of COLREGs. The temporary waypoint is calculated based on a real-time optimal orientation angle, which is determined by an optimization objective function for angle deviation. The three parts for local collision avoidance are determined by a finite state machine. In simulations, a USV with actual dynamic constraints runs successfully in a static and dynamic multiobstacle environment, and the results indicate the feasibility and validity of the proposed algorithm.

Stochastic Effects of 2D Random Arrays of Cylinders on Rarefied Gas Permeability Using the Lattice Boltzmann Method

A large set of 2D random arrays of circular cylinders is generated to perform a statistical study on rarefied gas flow through micro-porous media. The flow regimes in this work lie for Knudsen numbers (Kn) ranging from the continuum to the transition regimes. Arrays are built by randomly placing cylinders with constant diameter with a uniform distribution without overlapping, and are generated for three target porosities. Fluid flow is assumed to be incompressible and isothermal. A modified lattice Boltzmann model is adopted to account for discrete effects, with slip-velocity boundary conditions and a Kn-dependent multi-relaxation time collision operator. The apparent permeability is modeled with Darcy’s law with a Klinkenberg-type relationship and compared with existing correlations. Velocity fields highlight the increasing contribution of fluid flow through small pores with increasing Kn. Numerical results show that porous media randomness leads to an uncertainty on rarefied gas permeability calculation despite the same structural characteristics and may not strictly follow a specific correlation. The influence of a local collision operator based on a local Kn instead of a global one in the numerical model is also studied. Results show that the permeability in rarefied regimes undergoes significant deviation when applying the local collision operator compared to the global one. These differences could result from a more accurate capture of the pore-scale behavior with a local Kn. Thus, it emphasizes the sensitivity of the model and the apparent permeability calculation to the appropriate definition of Kn.

Shake, rattle, and help each other along

Active Matter In classical statistical mechanics, the deterministic dynamics of a many-body system are replaced by a probabilistic description. Chvykov et al. work toward a similar description for the nonequilibrium self-organization of collectives of active particles. In these systems, continuously input energy drives localized fluctuations, but larger-scale ordering can emerge, such as in the flight of a flock of birds. A key concept in their theory is the importance of rattling, whereby ordered patterns emerge through local collisions between neighbors at specific frequencies. The authors demonstrate this behavior using a set of flapping robots and produce related simulations of the robot behavior. Science , this issue p. [90][1] [1]: /lookup/doi/10.1126/science.abc6182

Analysis of critical condition for dry and wet growth icing on insulators

Icing on the insulators of transmission lines threats the safe operation of power system. Most of the existing researches focus on the discharge and flashover characteristics of ice-covered insulators. There is a lack of research on the growing process of insulator icing. In this paper, the critical condition for the dry and wet growth icing on insulators is taken as the research object. The calculation formulas of the freezing time of droplets and the collision interval time between adjacent water droplets are derived, based on which, the parametric characterization of different insulator icing types is proposed. By deriving the expression of the local collision efficiency β3, the distribution of water film on the insulator surface is obtained, which is used to distinguish the dry and wet growth icing on insulators. It is found that the critical condition of dry and wet growth icing is the condition that make the area covered with water film decrease to zero. With the decrease of temperature and droplets’ size, the area covered with water film decreases from the periphery of the droplet collision region to the centre, which is consistent to the observation test result on Xuefeng Mountain.

Review of Domain Wall Dynamics Engineering in Magnetic Microwires.

The influence of magnetic anisotropy, post-processing conditions, and defects on the domain wall (DW) dynamics of amorphous and nanocrystalline Fe-, Ni-, and Co-rich microwires with spontaneous and annealing-induced magnetic bistability has been thoroughly analyzed, with an emphasis placed on the influence of magnetoelastic, induced and magnetocrystalline anisotropies. Minimizing magnetoelastic anisotropy, either by the selection of a chemical composition with a low magnetostriction coefficient or by heat treatment, is an appropriate route for DW dynamics optimization in magnetic microwires. Stress-annealing allows further improvement of DW velocity and hence is a promising method for optimization of DW dynamics in magnetic microwires. The origin of current-driven DW propagation in annealing-induced magnetic bistability is attributed to magnetostatic interaction of outer domain shell with transverse magnetization orientation and inner axially magnetized core. The beneficial influence of the stress-annealing on DW dynamics has been explained considering that it allows increasing of the volume of outer domain shell with transverse magnetization orientation at the expense of decreasing the radius of inner axially magnetized core. Such transverse magnetic anisotropy can similarly affect the DW dynamics as the applied transverse magnetic field and hence is beneficial for DW dynamics optimization. Stress-annealing allows designing the magnetic anisotropy distribution more favorable for the DW dynamics improvement. Results on DW dynamics in various families of nanocrystalline microwires are provided. The role of saturation magnetization on DW mobility improvement is discussed. The DW shape, its correlation with the magnetic anisotropy constant and the microwire diameter, as well as manipulation of the DW shape by induced magnetic anisotropy are discussed. The engineering of DW propagation through local stress-annealing and DW collision is demonstrated.

Nanoscale Cinematography of Soft Matter System under Liquid-Phase TEM

ConspectusOne emergent theme in “soft matter” is to understand and manipulate the self-organization of synthetic materials and biological entities in space and time at the underexplored nanoscale. Encoded at this length scale can be a diversity of spatiotemporally fluctuating dynamics that are critical to function, from phase transition of nanoparticle self-assemblies as reconfigurable devices and morphology development of polymer membranes as separation layers for wastewater reclamation to the transformation of membrane proteins as the gatekeeper for mass and information flow in living cells. Extensive research efforts have thus been focused on resolving and understanding such dynamics that typically occur in a liquid medium. The proliferation of methods such as liquid-phase atomic force microscopy, cryogenic electron microscopy, and super-resolution optical microscopy has greatly expanded our knowledge in the structure or dynamics of soft matter at the nanoscale. However, these techniques do not offer direct real-space, real-time imaging of the structural and functional dynamics in a native liquid environment with nanometer resolution. This lack of experimental dataset also renders predictive modeling or computation difficult. As a result, how nanoscale morphology and interaction of the constituents affect the self-organization pathways or broadly collective structural evolution, such as interconversion among metastable states, as well as the involved energy measures remains poorly understood.In this Account, we present our recent efforts in adapting and using a nanoscopic cinematography method relatively new to the soft matter community, liquid-phase transmission electron microscopy (TEM), to study the self-organization pathways of nanoscale colloidal matter. Liquid-phase TEM has opened a new avenue to investigate materials chemistry questions, such as electrochemistry and catalysis, nanomaterial diffusion and growth, and nucleation of minerals and atomic crystals. Applying it to soft matter systems involves tackling complications, including the electron beam’s modification of nanoscale colloidal interaction and the substrate effect present in the liquid chamber, for both of which we highlight achievements of control. In addition, we discuss a series of first-time imaging of self-organization pathways of nanoparticle systems, accessible only by liquid-phase TEM. At low nanoparticle concentrations, chaining of nanoparticles occurs following quantitatively the kinetic laws of polymerization. This analogy originated from local collision and pairwise interaction, which can be directly mapped from trajectory sampling. At high nanoparticle concentrations, collective phase behaviors such as crystallization and coalescence are observed with single-particle resolution, allowing for the charting of phase coordinates and thermodynamic quantities based on statistical mechanics principles. We also discuss the general applicability of these methods. Lastly, toward taking live videos of organic soft matter at the nanoscale, we highlight recent instrumental developments, including machine learning based liquid-phase TEM video analysis to account for low signal-to-noise ratio data sets and low-dose electron tomography to resolve three-dimensional morphologies. We foresee that the examples, techniques, and understandings pinpoint the beginning of a paradigm shift in soft matter studies, where knowledge at the nanoscale can be derived from direct “seeing”.

Process planning for five-axis support free additive manufacturing

Traditionally Additive Manufacturing (AM) is a two-dimensional layer-by-layer material deposition process which requires building the support structures along with the build of the desired model. Removal of the support structures is costly and time-consuming, especially for metal parts. Using a five-axis deposition machine has the potential to build structures without the need for supports. However, there is a lack of automated process planning software to support the full use of five-axis machines. This paper introduces an automated method that allows reorienting the part during the build using a five-axis machine. The reorientations still allow the part to be built using traditional planar deposition but without the use of supports. This requires that the part be decomposed into sub volumes, such that each sub volume has its build direction and can be built with planar layers without support structures. This paper presents algorithms to determine the sub-volumes, their orientations, and sequence, which form the major components of the process plan for manufacturing. The process plan is generated by sequencing the decomposed volumes while ensuring a lack of local collision with previously deposited volumes. An added benefit of this automated process is the ability to evaluate the feasibility of building the part in a support free manner. This can provide feedback to the designer on the support free manufacturability of the part. Examples illustrating the methodology and establishing the viability of the decomposition strategy are presented to verify the effectiveness of the algorithms.

Characteristics of icing without icicle on insulators under natural conditions

Icing on insulators seriously threatens the safe and stable operation of transmission lines. Most of the existing research has focused on the flashover characteristics of ice-covered insulators, lacking attention to the icing process on insulators. In order to reveal the mechanism of insulator icing, and lay the foundation for further establishment of insulator ice flashover model, this paper studies the characteristics of insulator icing without icicle. The characteristics of icing processes including the collision and freezing of water droplets, the generation and flowing of water film are analysed. On this basis, a three-dimensional numerical calculation model of insulator icing without icicle is established by using boundary element method, and the expressions of the local icing weight and thickness on the insulator surface are derived through the droplet local collision efficiency and freezing efficiency. Moreover, the time-varying characteristics of icing parameters are simulated. And the related icing tests are conducted on two kinds of insulators with different structures at Xuefeng Mountain. The results show that the distribution of collided water droplets and water film on the insulator surface will change with the increasing ice weight. Under natural conditions, compared with the aerodynamic insulator LDQ-100, the insulator LXY-120 with a smaller skirt diameter and a larger inclination angle can have a faster icing rate. However, at the insulator edge, the insulator LDQ-100 with a thinner shed has a greater icing rate of thickness. Compared with the test results, the simulation error of the model in this paper is within 15%.

A Novel Double Layered Hybrid Multi-Robot Framework for Guidance and Navigation of Unmanned Surface Vehicles in a Practical Maritime Environment

Formation control and cooperative motion planning are two major research areas currently being used in multi robot motion planning and coordination. The current study proposes a hybrid framework for guidance and navigation of swarm of unmanned surface vehicles (USVs) by combining the key characteristics of formation control and cooperative motion planning. In this framework, two layers of offline planning and online planning are integrated and applied on a practical marine environment. In offline planning, an optimal path is generated from a constrained A* path planning approach, which is later smoothed using a spline. This optimal trajectory is fed as an input for the online planning where virtual target (VT) based multi-agent guidance framework is used to navigate the swarm of USVs. This VT approach combined with a potential theory based swarm aggregation technique provides a robust methodology of global and local collision avoidance based on known positions of the USVs. The combined approach is evaluated with the different number of USVs to understand the effectiveness of the approach from the perspective of practicality, safety and robustness.

Effect of Finger Sliding Direction on Tactile Perception, Friction and Dynamics

Human beings use certain styles when perceiving physical characteristics of objects with fingers. Sliding direction of finger contact is among the aforementioned styles. Thus, understanding the effect of sliding direction on tactile perception is essential. This study investigated the perception and identification of the roughness of sandpapers in two different sliding directions: proximal and distal. Furthermore, the corresponding tribological and dynamical properties of finger contact were examined using a custom setup. It was observed that the sliding direction and sliding speed influenced subjective judgment. The interaction between these two factors was also considered. A macroscopically continuous movement with a series of local collisions among the contact asperities enhanced the tactile perception in the proximal sliding. While in the distal direction, stick–slip occurred and interrupted the collision process, and reduced the tactile perception. The stick–slip phenomenon is modulated by various factors, including the external force and variations in the skin stiffness, which is influenced by the fingernail as well as the different friction properties in the two sliding directions.

Path planning of autonomous agricultural machineries in complex rural road

This study mainly investigates the path planning and collision avoidance problem of autonomous agricultural machineries in the complex rural road. First, the grid model is used to build the global map of machineries and the particle swarm optimisation algorithm is used to search the global path. A non-linear dynamic inertia weight is introduced to avoid the local lock-up problem of traditional particle swarm optimisation. Then the artificial potential energy method for addressing the local collision problem is used, which is the sum of the attraction of the target point and the repulsion of the obstacles and the road boundary force. To solve the problem of the slip and roll of autonomous machinery caused by complex roads, a machinery dynamic model considering road curvature and topographic inclination is established, and the model predictive control algorithm is used to track the planned path. The simulation results show that proposed method can make the autonomous agricultural machinery driving safely and smoothly.

Dynamics of seeded blobs under the influence of inelastic neutral interactions

Field-aligned filaments, the so-called blobs, born at the edge of the magnetically confined region of tokamaks propagate radially outward into the scrape-off layer (SOL) region that allows for a substantial population of neutral particles compared to the region of confinement. The electrons and ions constituting the blob undergo both elastic and inelastic collisions with the neutral particles, and the latter leads to sources and sinks of the blob density, momentum, and heat. The influence of the inelastic collisions with neutrals on the evolution of seeded blobs is investigated numerically by the nHESEL drift-fluid model through a series of discrete scans in interactions, active source terms, and blob plasma parameters. In light of the results, the potential influence of local inelastic collisions on the SOL density shoulder formation is discussed. It is found that density sources increase the blob compactness, which delays the blob dispersion and decreases the dispersion rate. Density sources or momentum sinks also influence the blob dynamics by increasing the vorticity layer around the perturbation, whereas the pressure sources/sinks only affect the blob dynamics marginally. The change to the vorticity structure leads, in most cases, to a decrease in the radial velocity of the blob center of mass, although, at high source rates, a radial acceleration of the blob center of mass is also observed. Density sources may, thus, contribute to shoulder formation not only by increasing the density locally but also by changing the filament dynamics.

Local motion simulation using deep reinforcement learning

AbstractTraditional local motion simulation focuses largely on avoiding collisions in the next frame. However, due to its lack of forward looking, the global trajectory of agents usually seems unreasonable. As a method of optimizing the overall reward, deep reinforcement learning (DRL) can better correct the problems that exist in the traditional local motion simulation method. In this article, we propose a local motion simulation method integrating optimal reciprocal collision avoidance (ORCA) and DRL, referred to as ORCA‐DRL. The main idea of ORCA‐DRL is to perform local collision avoidance detection via ORCA and smooth the trajectory at the same time via DRL. We use a deep neural network (DNN) as the state‐to‐action mapping function, where the state information is detected by virtual visual sensors and the action space includes two continuous spaces: speed and direction. To improve data utilization and speed up the training process, we use the proximal policy optimization based on the actor–critic (AC) framework to update the DNN parameters. Three scenes (circle, hallway, and crossing) are designed to evaluate the performance of ORCA‐DRL. The results reveal that, compared with the ORCA, our proposed ORCA‐DRL method can: (a) reduce the total number of frames, leading to less time for agents to reach their destination; and (b) effectively avoid local optima, evidenced by smoothed global trajectories.

Dynamics of a hollow cathode discharge in the frequency range of 1–500 kHz

We present results from 2D axisymmetric fluid simulations of the partially-ionized Xe gas in a 25-A cathode discharge, for a range of flow rates (5–20 sccm) in which transition between the so-called spot and plume modes has been observed in laboratory experiments. The simulations for the first time capture very well the characteristic rise of the peak-to-peak amplitude in the keeper voltage oscillations with decreasing flow rate—the so-called ‘plume margin’—thereby allowing for a closer interrogation of the processes that drive them. The oscillations are found to be due to plasma dynamics in the near-plume region that increase in amplitude as the flow rate is reduced. The cathode interior remains largely quiescent and changes in the emitter peak temperature do not exceed 2% across the entire range of flow rates examined. Along the cathode centerline the amplitude of the electron density oscillations peak at a location where the number density of the Xe atoms has diminished to its minimum value and the degree of ionization is close to one. The computed spectra reveal that most of the oscillation power resides in the frequency range of 100–300 kHz. These frequencies are much greater than the local ionization collision frequency ( < 10 kHz), and are associated with wavelengths of several centimeters. Such wavelengths are many orders of magnitude greater than λD. The waves are largely longitudinal with wave vector in the axial direction, and phase velocity ω/k > 10 km s−1, which is more than 6× greater than the ion acoustic speed. Though the computed and measured plume margins are in good agreement the measured spectra show that most of the power resides between ∼60 and 90 kHz, lower than the computed range (100–300 kHz), and exhibits a sharper distribution. It is argued this is due to the presence of a rotational mode in the experiment that is coupled with the longitudinal mode, yielding a complex 3D plasma motion that cannot be captured by the simulations due to the assumption of azimuthal axisymmetry, but that it is indeed the longitudinal plasma motion that drives the transition from spot to plume modes. An anomalous resistivity has been invoked in the simulations that is based on the formulations of Sagdeev and Galeev for the saturation of IAT which is long known to exist in these discharges. Though the simulations do not account for possible anomalous heating of the ions, the idealized model for the electrons is found to be sufficient in capturing the abovementioned dynamics, yielding a multiplication coefficient α of the anomalous collision frequency να = αωpe(Te/Ti)(ue/uTe) that is within a factor of two of the estimated theoretical value.

Video Synopsis Based on Attention Mechanism and Local Transparent Processing

The increased number of video cameras makes an explosive growth in the amount of captured video, especially the increase of millions of surveillance cameras that operate 24 hours a day. Since video browsing and retrieval is time consuming, while video synopsis is one of the most effective ways for browsing and indexing such video that enables the review of hours of video in just minutes. How to generate the video synopsis and preserve the essential activities in the original video is still a costly and labor-intensive and time-intensive work. This paper proposes an approach to generating video synopsis with complete foreground and clearer trajectory of moving objects. Firstly, the one-stage CNN-based object detecting has been employed in object extraction and classification. Then, combining integrating the attention-RetinaNet with Local Transparency-Handling Collision (LTHC) algorithm is given out which results in the trajectory combination optimization and makes the trajectory of the moving object more clearly. Finally, the experiments show that the useful video information is fully retained in the result video, the detection accuracy is improved by 4.87% and the compression ratio reaches 4.94, but the reduction of detection time is not obvious.

Active stress field and fault kinematics of the Greater Caucasus

This work contributes to depict the current seismicity, fault kinematics, and state of stress in the Greater Caucasus (territories of Georgia, Azerbaijan and Russia). We merged and homogenized data from different earthquake catalogues, relocated ~1000 seismic events, created a database of 366 selected focal mechanism solutions, 239 of which are new, and performed a formal stress inversion. Preferential alignments of crustal earthquake foci indicate that most seismic areas are located along the southern margin of the belt and in the north-eastern sector. This is consistent with the presence of dominant active WNW-ESE faults, parallel to the mountain range. In the entire Greater Caucasus, a dominant NNE-SSW-oriented greatest principal stress (σ1) controls the over-all occurrence of earthquakes of minor and major magnitude. Main earthquakes are characterized by a vertical least principal stress (σ3), corresponding to reverse kinematics. Reverse slip is more common along the southwestern and north-eastern foothills of the Greater Caucasus, although in these areas there are also scattered strike-slip events. This suggests the presence of local stress fields with horizontal σ1 and σ3. In the central-southern part of the mountain belt, in correspondence of the local collision between the Lesser and the Greater Caucasus, σ1 rotates to NNW-SSE. The strike-slip events, instead, dominate along the southern flank of the central-eastern mountain range; this is interpreted as the effect of the collision that promotes eastward escape of the tectonic blocks located to the east.

Control of the size of etchable ion tracks in PVDF – Irradiation in an oxygen atmosphere and with fullerene C60

We pursue the possibility of controlling the size of etchable ion tracks of poly(vinylidene-fluoride) (PVDF) films. At first, 330 MeV 40Ar ions were irradiated to a 25-µm-thick PVDF film in oxygen using the uniform-beam formation/irradiation system at Takasaki Ion Accelerators for Advanced Radiation Application (TIARA) cyclotron of National Institutes for Quantum and Radiological Science and Technology (QST, Japan). Compared with the irradiation in a high vacuum, which is the conventional method, it is clear that more oxygen-containing functional groups are formed on the surface, the size of the pores becomes larger and the track-etching rate is accelerated. The result shows that the irradiation in oxygen effectively induces oxidation of ion tracks and the oxidation widens the size of the etchable ion tracks. Next, we attempted the irradiation with 6-MeV fullerene C60+ ions. Compared with the irradiation with 100-keV C+ monatomic ions of the same velocity, the fullerene-ion irradiation induced large sized etchable tracks on the surface of PVDF. The result could represent the effect of local and simultaneous collisions by the aggregated ions. Both of the irradiations showed widening the size of etchable ion tracks and we were able to control the pore diameter by varying the etching time.

Unified gas-kinetic wave-particle methods. II. Multiscale simulation on unstructured mesh

In this paper, we present a unified gas-kinetic wave-particle (UGKWP) method on unstructured mesh for the multiscale simulation of continuum and rarefied flow. Inheriting from the multiscale transport in the unified gas-kinetic scheme (UGKS), the integral solution of the kinetic model equation is employed in the construction of the UGKWP method to model the flow physics on the scales of cell size and time step. A novel wave-particle adaptive formulation is introduced in the UGKWP method to describe the flow dynamics in each control volume. The local gas evolution is constructed through the dynamical interaction of the deterministic hydrodynamic wave and the stochastic kinetic particle. To model the gas dynamics on the scales of cell size and time step, the decomposition, interaction, and evolution of the hydrodynamic wave and the kinetic particle depend on the ratio of time step to local collision time. In the rarefied flow regime, the UGKWP method recovers the nonequilibrium flow physics by discrete particles and performs as a stochastic particle method. In the continuum flow regime, the UGKWP method captures the flow behavior solely by macroscopic variable evolution and becomes a gas-kinetic hydrodynamic flow solver, the same as the gas-kinetic scheme, for viscous and heat-conducting Navier–Stokes solutions. In the transition regime, both kinetic particle and hydrodynamic wave contribute adaptively in the UGKWP to capture the peculiar nonequilibrium flow physics in a most efficient way. In different flow regimes, the Sod shock tube, lid-driven cavity flow, laminar boundary layer, and high-speed flow around a circular cylinder are computed to validate the UGKWP method on unstructured mesh. The UGKWP method obtains the same UGKS solutions in all Knudsen regimes. However, with an automatic wave-particle decomposition, the UGKWP method becomes very efficient. For example, at Mach number 30 and Knudsen number 0.1, the UGKWP has several-order-of-magnitude reductions in computational cost and memory requirement in comparison with UGKS. Overall, the UGKWP can capture the gas dynamics in all flow regimes efficiently and accurately from the free molecular transport to the Navier-Stokes flow evolution.

Collisions on local roads: model development and policy level scenario analysis

The problem of collisions on local roads has received little specific attention despite the considerable number of such collisions that occur each year. First part of this study identifies the factors that influence local road collision frequency at traffic analysis zone (TAZ) level with a particular focus on the planning and policy related variables. The City of Calgary is used as a case study, where we focus on the impacts of land use, demographic characteristics, and travel characteristics. We also investigate the effects of some key transportation planning parameters for which there have been very limited studies, including the number of personal and commercial trips and the employment numbers in various categories. This study examines the impact of the number of trips made by automobile versus more sustainable transport modes like transit, walking, and biking for personal travel. It also examines the impact of commercial truck movement on the number of collisions on local roads in a TAZ. The impact of transit-oriented development zone initiatives is explored, as is the relationship between the predominant land use type (e.g., residential, commercial, industrial) and the number of collisions on local roads. In the second part, collision prediction models were linked with regional transportation model (RTM), which is calibrated and modeled in EMME. Since the choice of transportation mode is explicitly modeled through utility functions in the RTM, the proposed approach will allow us to do scenario analysis for planning and policy level issues proactively such as impact on local collisions due to change in fuel price, parking cost, transit headway, and transit fare. Results showed that property damage only (PDO) and fatal and injury (FI) collisions decreased by 13% and 6%, respectively, when fuel price was doubled. It was also observed that PDO and FI collisions decreased by 8% and 5%, respectively, when parking cost was doubled. PDO and FI collisions decreased by 7% and 4%, respectively, when transit headway was reduced to half. When transit fare was reduced to half, PDO and FI collisions decreased by 5% and 2%, respectively. PDO and FI collisions decreased by 10% and 5%, respectively, when transit fare was set to zero. These scenario analyses demonstrate how the impact of transportation planning or policy level issues on the collision count on local roads can be incorporated in our proposed model.

Safe Navigation With Human Instructions in Complex Scenes

In this letter, we present a robotic navigation algorithm with natural language interfaces that enables a robot to safely walk through a changing environment with moving persons by following human instructions such as “go to the restaurant and keep away from people.” We first classify human instructions into three types: goal, constraints, and uninformative phrases. Next, we provide grounding in a dynamic manner for the extracted goal and constraint items along with the navigation process to deal with target objects that are too far away for sensor observation and the appearance of moving obstacles such as humans. In particular, for a goal phrase (e.g., “go to the restaurant”), we ground it to a location in a predefined semantic map and treat it as a goal for a global motion planner, which plans a collision-free path in the workspace for the robot to follow. For a constraint phrase (e.g., “keep away from people”), we dynamically add the corresponding constraint into a local planner by adjusting the values of a local costmap according to the results returned by the object detection module. The updated costmap is then used to compute a local collision avoidance control for the safe navigation of the robot. By combining natural language processing, motion planning, and computer vision, our developed system can successfully follow natural language navigation instructions to achieve navigation tasks in both simulated and real-world scenarios. Videos are available at https://sites.google.com/view/snhi .