Struct TrajectoryRecorder 

pub struct TrajectoryRecorder {
    pub q_history: Vec<Vec<f64>>,
    pub v_history: Vec<Vec<f64>>,
    pub ctrl_history: Vec<Vec<f64>>,
    pub time_history: Vec<f64>,
    pub sensor_history: Vec<Vec<SensorOutput>>,
}

Records trajectories of (q, v, ctrl, time) for ML training.

Fields

q_history: Vec<Vec<f64>>

Recorded positions at each timestep.

v_history: Vec<Vec<f64>>

Recorded velocities at each timestep.

ctrl_history: Vec<Vec<f64>>

Recorded control inputs at each timestep.

time_history: Vec<f64>

Timestamps for each step.

sensor_history: Vec<Vec<SensorOutput>>

Sensor outputs at each timestep.

Implementations

impl TrajectoryRecorder

pub fn new() -> Self

Create a new empty trajectory recorder.

pub fn record(&mut self, state: &State)

Record the current state.

pub fn record_with_sensors(&mut self, state: &State, sensors: Vec<SensorOutput>)

Record state with sensor outputs.

pub fn len(&self) -> usize

Number of timesteps recorded.

pub fn is_empty(&self) -> bool

Check if recorder is empty.

pub fn clear(&mut self)

Clear all recorded data.

pub fn to_json(&self) -> Result<String, Error>

Export to JSON string.

pub fn to_json_file(&self, path: &str) -> Result<(), Box<dyn Error>>

Export to JSON file.

pub fn to_flat_dict(&self) -> HashMap<String, Vec<f64>>

Convert to a dictionary-like structure suitable for numpy/ML frameworks.

Returns:

• “q”: (nsteps, nq) flattened data
• “v”: (nsteps, nv) flattened data
• “ctrl”: (nsteps, nv) flattened data
• “time”: (nsteps,) flattened data

pub fn stats(&self) -> TrajectoryStats

Get trajectory statistics.

Trait Implementations

impl Default for TrajectoryRecorder

fn default() -> Self

Returns the “default value” for a type.