船舶航速预测数据

通过预测船舶的航速，可以更有效地规划船舶的航行路径，避免与其他船舶发生碰撞。可以更好地调配港口资源，如泊位、装卸设备等，减少等待时间，提升港口的运营效率。通过船舶自带的AIS系统自动识别和传输船舶的静态信息，包括船名和动态信息（如位置、航向、航速等），对收集的AIS数据进行清洗等预处理操作，以提高模型的训练效率和预测精度。使用深度学习框架构建RNN模型，使用LSTM来处理序列数据，隐藏层的计算是递归的，即每个时间步的隐藏状态都依赖于前一个时间步的隐藏状态和当前时间步的输入。使用RNN模型以时间戳、速度、角度为输入特征，基于 RNN的船舶航速预测模型公式：Yt+1 =f({Xt-9 ,Xt-8 ,…,Xt-1 ,Xt})，f 是RNN模型映射函数，Yt+1 是预测的下一个时间点的航速。RNN的输入为船舶在过去10个时间返回的速度和角度，记作Xt-9 ,Xt-8 ,…,Xt-1 ,X，其中每个Xi是一个包含速度和角度的二维向量。

Predicting ship speed enables more effective vessel route planning, collision avoidance with other ships, optimized allocation of port resources such as berths and loading/unloading equipment, reduced waiting time, and improved port operational efficiency. Static information including vessel names and dynamic information such as position, course, speed, etc. of ships are automatically identified and transmitted via the shipborne Automatic Identification System (AIS). Collected AIS data undergoes preprocessing steps including data cleaning to enhance model training efficiency and prediction accuracy. A Recurrent Neural Network (RNN) model is constructed using a deep learning framework, with Long Short-Term Memory (LSTM) employed to handle sequential data. The hidden layer computation is recursive, meaning the hidden state at each time step depends on the hidden state of the preceding time step and the input at the current time step. The RNN model takes timestamps, speed and angle as input features. The formula for the RNN-based ship speed prediction model is: $Y_{t+1} = f({X_{t-9}, X_{t-8}, dots, X_{t-1}, X_t})$, where $f$ is the mapping function of the RNN model, and $Y_{t+1}$ is the predicted ship speed at the next time step. The input to the RNN consists of the speed and angle data from the past 10 time steps of the ship, denoted as $X_{t-9}, X_{t-8}, dots, X_{t-1}, X_t$, where each $X_i$ is a 2-dimensional vector containing speed and angle.

搜集汇总

数据集介绍

特点

船舶航速预测数据集由中创海洋科技股份有限公司提供，包含50399条船舶的静态和动态信息，每日更新。数据通过AIS系统收集，并使用RNN模型进行航速预测，应用于航行路径规划和港口资源调配。