Goal location prediction based on deep learning using RGB-D camera

In the navigation system, the desired destination position plays an essential role since the path planning algorithms takes a current location and goal location as inputs as well as the map of the surrounding environment. The generated path from path planning algorithm is used to guide a user to his final destination. This paper presents a proposed algorithm based on RGB-D camera to predict the goal coordinates in 2D occupancy grid map for visually impaired people navigation system. In recent years, deep learning methods have been used in many object detection tasks. So, the object detection method based on convolution neural network method is adopted in the proposed algorithm. The measuring distance between the current position of a sensor and the detected object depends on the depth data that is acquired from RGB-D camera. Both of the object detected coordinates and depth data has been integrated to get an accurate goal location in a 2D map. This proposed algorithm has been tested on various real-time scenarios. The experiments results indicate to the effectiveness of the proposed algorithm.

Hippocampal place cells use vector computations to navigate

One function of the Hippocampal Cognitive Map is to provide information about salient locations in familiar environments such as those containing reward or danger, and to support navigation towards or away from those locations. Although much is known about how the hippocampus encodes location in world-centred coordinates, how it supports flexible navigation is less well understood. We recorded from CA1 place cells while rats navigated to a goal or freely foraged on the honeycomb maze. The maze tests the animal's ability to navigate using indirect as well as direct paths to the goal and allows the directionality of place cells to be assessed at each choice point during traversal to the goal. Place fields showed strong directional polarization in the navigation task, and to a lesser extent during random foraging. This polarization was characterized by vector fields which converged to sinks distributed throughout the environment. The distribution of these convergence sinks was centred near the goal location, and the population vector field converged on the goal, providing a strong navigational signal. Changing the goal location led to the movement of ConSinks and vector fields towards the new goal and within-days, the ConSink distance to the goal decreased with continued training. The honeycomb maze allows the independent assessment of spatial representation and spatial action in place cell activity and shows how the latter depends on the former. The results suggest a vector-based model of how the hippocampus supports flexible navigation, allowing animals to select optimal paths to destinations from any location in the environment.

The effects of spatial stability and cue type on spatial learning: Implications for theories of parallel memory systems

Some theories of spatial learning predict that associative rules apply under only limited circumstances. For example, learning based on a boundary has been claimed to be immune to cue competition effects because boundary information is the basis for the formation of a cognitive map, whilst landmark learning does not involve cognitive mapping. This is referred to as the cue type hypothesis. However, it has also been claimed that cue stability is a prerequisite for the formation of a cognitive map, meaning that whichever cue type was perceived as stable would enter a cognitive map and thus be immune to cue competition, while unstable cues will be subject to cue competition, regardless of cue type. In experiments 1 and 2 we manipulated the stability of boundary and landmark cues when learning the location of two hidden goals. One goal location was constant with respect to the boundary, and the other constant with respect to the landmark cues. For both cue types, the presence of distal orientation cues provided directional information. For half the participants the landmark cues were unstable relative to the boundary and orientation cues, whereas for the remainder of the participants the boundary was unstable relative to landmarks and orientation cues. In a second stage of training, all cues remained stable so that both goal locations could be learned with respect to both landmark and boundary information. According to the cue type hypothesis, boundary information should block learning about landmarks regardless of cue stability. According to the cue stability hypothesis, however, landmarks should block learning about the boundary when the landmarks appear stable relative to the boundary. Regardless of cue type or stability the results showed reciprocal blocking, contrary to both formulations of incidental cognitive mapping. Experiment 3 established that the results of Experiments 1 and 2 could not be explained in terms of difficulty in learning certain locations with respect to different cue types. In a final experiment, following training in which both landmarks and boundary cues signalled two goal locations, a new goal location was established with respect to the landmark cues, before testing with the boundary, which had never been used to define the new goal location. The results of this novel test of the interaction between boundary and landmark cues indicated that new learning with respect to the landmark had a profound effect on navigation with respect to the boundary, counter to the predictions of incidental cognitive mapping of boundaries.

EXPRESS: Cue Combination in Goal-Oriented Navigation

This study examined cue combination of self-motion and landmark cues in goal-localization. In an immersive virtual environment, before walking a two-leg path, participants learned the locations of three goal objects (one at the path origin, i.e., home) and landmarks. After walking the path without seeing landmarks or goals, participants indicated the locations of the home and non-home goals in four conditions: 1) path integration only, 2) landmarks only, 3) both path integration and the landmarks, and 4) path integration and rotated landmarks. The ratio of the length between the testing position (P) and the turning point (T) over the length between the T and the three goals (G) (i.e., PT/TG) was manipulated. The results showed the cue combination consistently for participants’ heading estimates but not for goal-localization. In Experiments 1-2 (using distal landmarks), the cue combination for goal estimates appeared in a small length ratio (PT/TG=0.5) but disappeared in a large length ratio (PT/TG=2). In Experiments 3-4 (using proximal landmarks), while the cue combination disappeared for the home with a medium length ratio (PT/TG=1), it appeared for the non-home goal with a large length ratio (PT/TG=2) and only disappeared with a very large length ratio (PT/TG=3). These findings are explained by a model stipulating that cue combination occurs in self-localization (e.g., heading estimates), which leads to one estimate of the goal location; proximal landmarks produce another goal location estimate; these two goal estimates are then combined, which may only occur for non-home goals.

Impaired spatial learning and suppression of sharp wave ripples by cholinergic activation at the goal location

The hippocampus plays a central role in long-term memory formation, and different hippocampal network states are thought to have different functions in this process. These network states are controlled by neuromodulatory inputs, including the cholinergic input from the medial septum. Here, we used optogenetic stimulation of septal cholinergic neurons to understand how cholinergic activity affects different stages of spatial memory formation in a reward-based navigation task in mice. We found that optogenetic stimulation of septal cholinergic neurons (1) impaired memory formation when activated at goal location but not during navigation; (2) reduced sharp wave-ripple (SWR) incidence at goal location; and (3) reduced SWR incidence and enhanced theta-gamma oscillations during sleep. These results underscore the importance of appropriate timing of cholinergic input in long-term memory formation, which might help explain the limited success of cholinesterase inhibitor drugs in treating memory impairment in Alzheimer's disease.

Towards Energy-Aware Feedback Planning for Long-Range Autonomous Underwater Vehicles

Ocean ecosystems have spatiotemporal variability and dynamic complexity that require a long-term deployment of an autonomous underwater vehicle for data collection. A new generation of long-range autonomous underwater vehicles (LRAUVs), such as the Slocum glider and Tethys-class AUV, has emerged with high endurance, long-range, and energy-aware capabilities. These new vehicles provide an effective solution to study different oceanic phenomena across multiple spatial and temporal scales. For these vehicles, the ocean environment has forces and moments from changing water currents which are generally on the order of magnitude of the operational vehicle velocity. Therefore, it is not practical to generate a simple trajectory from an initial location to a goal location in an uncertain ocean, as the vehicle can deviate significantly from the prescribed trajectory due to disturbances resulted from water currents. Since state estimation remains challenging in underwater conditions, feedback planning must incorporate state uncertainty that can be framed into a stochastic energy-aware path planning problem. This article presents an energy-aware feedback planning method for an LRAUV utilizing its kinematic model in an underwater environment under motion and sensor uncertainties. Our method uses ocean dynamics from a predictive ocean model to understand the water flow pattern and introduces a goal-constrained belief space to make the feedback plan synthesis computationally tractable. Energy-aware feedback plans for different water current layers are synthesized through sampling and ocean dynamics. The synthesized feedback plans provide strategies for the vehicle that drive it from an environment’s initial location toward the goal location. We validate our method through extensive simulations involving the Tethys vehicle’s kinematic model and incorporating actual ocean model prediction data.

Encoding of environmental cues in central amygdala neurons during foraging

SUMMARYIn order to successfully forage in an environment filled with rewards and threats, animals need to rely on familiar structures of their environment that signal food availability. The central amygdala (CeA) is known to mediate a panoply of consummatory and defensive behaviors, yet how specific activity patterns within CeA subpopulations guide optimal choices is incompletely understood. In a paradigm of appetitive conditioning in which mice freely forage for food across a continuum of cues, we find that two major subpopulations of CeA neurons, Somatostatin-positive (CeASst) and protein kinase Cδ-positive (CeAPKCδ) neurons can assign motivational properties to environmental cues and encode memory of goal location. While the proportion of food responsive cells was higher within CeASst than CeAPKCδ neurons, only the activities of CeAPKCδ, but not CeASst, neurons were required for learning of contextual food cues. Since CeAPKCδ neurons are known to promote a range of defensive behaviors, our findings point to a model in which CeA circuit components are not organized in specialized functional units but can process both aversive and rewarding information in a context and experience dependent manner.HIGHLIGHTSTwo populations of central amygdala (CeA) neurons, CeAPKCδ and CeASst neurons can assign motivational properties to environmental cues and encode memory of goal location.The proportion of food responsive cells was higher among CeASst, than CeAPKCδ neurons.The activities of CeAPKCδ, but not CeASst, neurons are required for learning of contextual food cues. CeAPKCδ neurons represent a “general encoding” population selecting defensive and appetitive responses depending on context.

Gaze Behaviour Reveals Flexible Encoding of Competing Reach Goals Under Conditions of Target Uncertainty

In daily tasks, we are often confronted with competing potential targets and must select one to act on. It has been suggested that, prior to target selection, the human brain encodes the motor goals of multiple, potential targets. However, this view remains controversial and it has been argued that only a single motor goal is encoded, or that motor goals are only specified after target selection. To investigate this issue, we measured participants’ gaze behaviour while viewing two potential reach targets, one of which was cued after a preview period. We applied visuomotor rotations to dissociate each visual target location from its corresponding motor goal location; i.e., the location participants needed to aim their hand toward to bring the rotated cursor to the target. During the preview period, participants most often fixated both motor goals but also frequently fixated one, or neither, motor goal location. Further gaze analysis revealed that on trials in which both motor goals were fixated, both locations were held in memory simultaneously. These findings show that, at the level of single trials, the brain most often encodes multiple motor goals prior to target selection, but may also encode either one or no motor goals. This result may help reconcile a key debate concerning the specification of motor goals in cases of target uncertainty.

Enhancement of Hippocampal Spatial Decoding Using a Dynamic Q-Learning Method With a Relative Reward Using Theta Phase Precession

Hippocampal place cells and interneurons in mammals have stable place fields and theta phase precession profiles that encode spatial environmental information. Hippocampal CA1 neurons can represent the animal’s location and prospective information about the goal location. Reinforcement learning (RL) algorithms such as Q-learning have been used to build the navigation models. However, the traditional Q-learning ([Formula: see text]Q-learning) limits the reward function once the animals arrive at the goal location, leading to unsatisfactory location accuracy and convergence rates. Therefore, we proposed a revised version of the Q-learning algorithm, dynamical Q-learning ([Formula: see text]Q-learning), which assigns the reward function adaptively to improve the decoding performance. Firing rate was the input of the neural network of [Formula: see text]Q-learning and was used to predict the movement direction. On the other hand, phase precession was the input of the reward function to update the weights of [Formula: see text]Q-learning. Trajectory predictions using [Formula: see text]Q- and [Formula: see text]Q-learning were compared by the root mean squared error (RMSE) between the actual and predicted rat trajectories. Using [Formula: see text]Q-learning, significantly higher prediction accuracy and faster convergence rate were obtained compared with [Formula: see text]Q-learning in all cell types. Moreover, combining place cells and interneurons with theta phase precession improved the convergence rate and prediction accuracy. The proposed [Formula: see text]Q-learning algorithm is a quick and more accurate method to perform trajectory reconstruction and prediction.

Impaired spatial learning and suppression of sharp wave ripples by cholinergic activation at the goal location

The hippocampus plays a central role in long-term memory formation, and different hippocampal network states are thought to have different functions in this process. These network states are controlled by neuromodulatory inputs, including the cholinergic input from the medial septum. Here, we used optogenetic stimulation of septal cholinergic neurons to understand how cholinergic activity affects different stages of spatial memory formation in a reward-based navigation task in mice. We found that optogenetic stimulation of septal cholinergic neurons (1) impaired memory formation when activated at goal location but not during navigation; (2) reduced sharp wave-ripple (SWR) incidence at goal location; and (3) reduced SWR incidence and enhanced theta-gamma oscillations during sleep. These results underscore the importance of appropriate timing of cholinergic input in long-term memory formation, which might help explain the limited success of cholinesterase inhibitor drugs in treating memory impairment in Alzheimer's disease.