test_acceleration_control_planner.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import theseus as th\n",
    "from objective.acceleration_based import AccelerationBasedObjective\n",
    "import torch\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "dtype = torch.double\n",
    "device = torch.device(\"cpu\")\n",
    "# device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "CELL_SIZE = 0.05\n",
    "HORIZON = 30\n",
    "TOTAL_TIME = 60.0 # this dictates the discretization time step (dt = TOTAL_TIME / HORIZON)\n",
    "CURRENT_STATE = torch.tensor((45*CELL_SIZE, 40*CELL_SIZE, 0))  # x, y, theta\n",
    "GOAL_POSITION = torch.tensor((45*CELL_SIZE, 20*CELL_SIZE))  # x, y\n",
    "SDF_ORIGIN = (0, 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "objective = AccelerationBasedObjective(\n",
    "    total_time=TOTAL_TIME,\n",
    "    horizon=HORIZON,\n",
    "    x_velocity_bounds=torch.tensor([-5, 5], dtype=dtype),\n",
    "    y_velocity_bounds=torch.tensor([-5, 5], dtype=dtype),\n",
    "    x_acceleration_bounds=torch.tensor([-1, 1], dtype=dtype),\n",
    "    y_acceleration_bounds=torch.tensor([-1, 1], dtype=dtype),\n",
    "    robot_radius=1,  # REPLACE WITH ACTUAL RADIUS\n",
    "    safety_distance=1,  # REPLACE WITH ACTUAL SAFETY DISTANCE\n",
    "    local_map_size=60,  # REPLACE WITH ACTUAL MAP SIZE\n",
    "    # goal_cost = 50,\n",
    "    # quadratic_acceleration_cost = 10,\n",
    "    # velocity_bounds_cost = 2000,\n",
    "    # acceleration_bounds_cost = 2000,\n",
    "    # current_state_cost = 2000,\n",
    "    # dynamic_cost = 2000,\n",
    "    # collision_cost = 2000\n",
    ")\n",
    "\n",
    "optimizer = th.LevenbergMarquardt(\n",
    "    objective,\n",
    "    th.CholeskyDenseSolver,\n",
    "    max_iterations=15,\n",
    "    step_size=1,\n",
    ")\n",
    "motion_planner = th.TheseusLayer(optimizer)\n",
    "motion_planner.to(device, dtype=torch.double)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_inputs(horizon):\n",
    "    # Returns a dictionary with state and acceleration variable names associated to a\n",
    "    # straight line trajectory between start and goal in SE(2) space\n",
    "\n",
    "    # Calculate start and goal positions\n",
    "    # start_pos = start[..., :2]  # Extracting the 2D position from SE(2)\n",
    "    # goal_pos = goal  # Assuming goal is already a 2D point\n",
    "\n",
    "    # # Calculate distances and average accelerations for each batch element\n",
    "    # start_goal_dist = goal_pos - start_pos  # Shape: (batch_size, 2)\n",
    "    # avg_acc = 2 * start_goal_dist / (total_time**2)  # Shape: (batch_size, 2)\n",
    "    # unit_horizon = start_goal_dist / (horizon - 1)  # Shape: (batch_size, 2)\n",
    "\n",
    "    # Create input dictionary\n",
    "    input_dict = {}\n",
    "\n",
    "    for i in range(horizon + 1):\n",
    "        # state_i = start_pos + unit_horizon * i  # Shape: (batch_size, 2)\n",
    "        # Concatenate the 2D position with the orientation\n",
    "        input_dict[f\"state_{i}\"] = torch.tensor([0.0, 0.0, 1.0, 0.0], requires_grad=True, dtype=torch.double).view(1, 4)\n",
    "\n",
    "        if i == horizon:\n",
    "            continue\n",
    "\n",
    "        # if i == 0 or i == horizon - 1:\n",
    "        #     # Zero acceleration at start and end\n",
    "        acceleration_i = torch.zeros(1, 2, requires_grad = True, dtype=torch.double)\n",
    "\n",
    "        input_dict[f\"acceleration_{i}\"] = acceleration_i\n",
    "\n",
    "    return input_dict\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([60, 60])\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "file_path = \"sdf/sdf_array_2.npy\"\n",
    "sdf = torch.tensor(np.load(file_path), dtype=dtype)\n",
    "plt.imshow(sdf)\n",
    "plt.colorbar()\n",
    "print(sdf.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "planner_inputs = {\n",
    "    # REPLACE WITH SDF DATA\n",
    "    'sdf_data': sdf.unsqueeze(0).to(device),\n",
    "    'sdf_origin': torch.tensor([SDF_ORIGIN[0], SDF_ORIGIN[1]], dtype=dtype).unsqueeze(0).to(device),\n",
    "    'cell_size': torch.tensor([CELL_SIZE], dtype=dtype).unsqueeze(0).to(device),\n",
    "    'current_state': torch.tensor([CURRENT_STATE[0], CURRENT_STATE[1], torch.cos(CURRENT_STATE[2]), torch.sin(CURRENT_STATE[2])], dtype=dtype).unsqueeze(0).to(device),\n",
    "    'goal_position': torch.tensor([GOAL_POSITION[0], GOAL_POSITION[1]], dtype=dtype).unsqueeze(0).to(device),\n",
    "}\n",
    "planner_inputs.update(get_inputs(HORIZON))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Nonlinear optimizer. Iteration: 0. Error: 18352343.75\n",
      "Nonlinear optimizer. Iteration: 1. Error: 33302940.954343528\n",
      "Nonlinear optimizer. Iteration: 2. Error: 27842818.682020523\n",
      "Nonlinear optimizer. Iteration: 3. Error: 18217325.51924601\n",
      "Nonlinear optimizer. Iteration: 4. Error: 371006.6598314684\n",
      "Nonlinear optimizer. Iteration: 5. Error: 31513.474972374308\n",
      "Nonlinear optimizer. Iteration: 6. Error: 150450.168593053\n",
      "Nonlinear optimizer. Iteration: 7. Error: 5060.1353712476075\n",
      "Nonlinear optimizer. Iteration: 8. Error: 5060.134743922179\n",
      "Nonlinear optimizer. Iteration: 9. Error: 5060.134743912179\n",
      "Time: 0.9175612259996342\n"
     ]
    }
   ],
   "source": [
    "import time\n",
    "start = time.perf_counter()\n",
    "# with torch.no_grad():\n",
    "final_values, info = motion_planner.forward(\n",
    "    planner_inputs,\n",
    "    optimizer_kwargs={\n",
    "        \"track_best_solution\": True,\n",
    "        \"verbose\": True,\n",
    "        \"damping\": 0.1,\n",
    "    }\n",
    ")\n",
    "end = time.perf_counter()\n",
    "print(f\"Time: {end - start}\")\n",
    "\n",
    "solution_dictionary = info.best_solution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def dummy_loss(solution_dict):\n",
    "    loss = 0.0\n",
    "    for i in range(HORIZON + 1):\n",
    "        state_key = f\"state_{i}\"\n",
    "        state = solution_dict[state_key]\n",
    "        loss += torch.sum(state)\n",
    "\n",
    "    for i in range(HORIZON):\n",
    "        acceleration_key = f\"acceleration_{i}\"\n",
    "        acceleration = solution_dict[acceleration_key]\n",
    "        loss += torch.sum(acceleration)\n",
    "\n",
    "    return loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[tensor([[-7.3266e-15,  5.8466e-15]], dtype=torch.float64), tensor([[-7.1083e-15,  5.6963e-15]], dtype=torch.float64), tensor([[-6.5939e-15,  5.3015e-15]], dtype=torch.float64), tensor([[-5.7469e-15,  4.6331e-15]], dtype=torch.float64), tensor([[-4.5139e-15,  3.6483e-15]], dtype=torch.float64), tensor([[-2.8268e-15,  2.2927e-15]], dtype=torch.float64), tensor([[-6.0581e-16,  5.0223e-16]], dtype=torch.float64), tensor([[ 1.0348e-15, -8.2097e-16]], dtype=torch.float64), tensor([[ 2.1834e-15, -1.7478e-15]], dtype=torch.float64), tensor([[ 2.9251e-15, -2.3470e-15]], dtype=torch.float64), tensor([[ 3.3394e-15, -2.6823e-15]], dtype=torch.float64), tensor([[ 3.4982e-15, -2.8117e-15]], dtype=torch.float64), tensor([[ 3.4650e-15, -2.7862e-15]], dtype=torch.float64), tensor([[ 3.2940e-15, -2.6497e-15]], dtype=torch.float64), tensor([[ 3.0312e-15, -2.4389e-15]], dtype=torch.float64), tensor([[ 2.7137e-15, -2.1841e-15]], dtype=torch.float64), tensor([[ 2.3715e-15, -1.9091e-15]], dtype=torch.float64), tensor([[ 2.0274e-15, -1.6325e-15]], dtype=torch.float64), tensor([[ 1.6986e-15, -1.3680e-15]], dtype=torch.float64), tensor([[ 1.3969e-15, -1.1253e-15]], dtype=torch.float64), tensor([[ 1.1302e-15, -9.1075e-16]], dtype=torch.float64), tensor([[ 9.0277e-16, -7.2770e-16]], dtype=torch.float64), tensor([[ 7.1606e-16, -5.7740e-16]], dtype=torch.float64), tensor([[ 5.6935e-16, -4.5928e-16]], dtype=torch.float64), tensor([[ 4.6015e-16, -3.7135e-16]], dtype=torch.float64), tensor([[ 3.8459e-16, -3.1048e-16]], dtype=torch.float64), tensor([[ 3.3764e-16, -2.7266e-16]], dtype=torch.float64), tensor([[ 3.1336e-16, -2.5309e-16]], dtype=torch.float64), tensor([[ 3.0498e-16, -2.4632e-16]], dtype=torch.float64), tensor([[ 3.0494e-16, -2.4628e-16]], dtype=torch.float64)]\n"
     ]
    }
   ],
   "source": [
    "loss = dummy_loss({k: t.tensor.clone() for (k, t) in motion_planner.objective.optim_vars.items()})\n",
    "loss.backward()\n",
    "print([planner_inputs[f'acceleration_{i}'].grad for i in range(HORIZON)])\n",
    "# NOTE: only the initial acceleration tensors actually have gradients, since those are the only leaf tensors. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "state_values = {\n",
    "    key: value for key, value in solution_dictionary.items() if \"state\" in key\n",
    "}\n",
    "trajectory = torch.stack(list(state_values.values()), dim=0)\n",
    "\n",
    "plt.imshow(sdf)\n",
    "for x, y, a, b in trajectory.squeeze():\n",
    "    plt.plot(x/CELL_SIZE, y/CELL_SIZE, 'bo')\n",
    "\n",
    "plt.plot(CURRENT_STATE[0]/CELL_SIZE, CURRENT_STATE[1]/CELL_SIZE, 'rx')\n",
    "plt.plot(GOAL_POSITION[0]/CELL_SIZE, GOAL_POSITION[1]/CELL_SIZE, 'gx')\n",
    "plt.colorbar()\n",
    "plt.show()\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "theseus",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}