Coverage for src/gncpy/dynamics/basic/curvilinear_motion.py: 73%

49 statements  

« prev     ^ index     » next       coverage.py v7.2.7, created at 2023-07-19 05:48 +0000

1import numpy as np 

2 

3from warnings import warn 

4from .nonlinear_dynamics_base import NonlinearDynamicsBase 

5 

6 

7 

8class CurvilinearMotion(NonlinearDynamicsBase): 

9 r"""Implements general curvilinear motion model in 2d. 

10 

11 This is a slight variation from normal :class:`.NonlinearDynamicsBase` classes 

12 because it does not use a list of continuous functions but instead has 

13 the state and input matrices coded directly. As a result, it also does not 

14 use the control model attribute because it is hardcoded in the 

15 :meth:`.get_input_mat` function. Also, the angle state should be kept between 0-360 

16 degrees. 

17 

18 Notes 

19 ----- 

20 This implements the following system of ODEs. 

21 

22 .. math:: 

23 

24 \begin{align} 

25 \dot{x} &= v cos(\psi) \\ 

26 \dot{y} &= v sin(\psi) \\ 

27 \dot{v} &= u_0 \\ 

28 \dot{\psi} &= u_1 

29 \end{align} 

30 

31 See :cite:`Li2000_SurveyofManeuveringTargetTrackingDynamicModels` for 

32 details. 

33 """ 

34 

35 state_names = ( 

36 "x pos", 

37 "y pos", 

38 "speed", 

39 "turn angle", 

40 ) 

41 

42 def __init__(self, **kwargs): 

43 """Initialize an object. 

44 

45 Parameters 

46 ---------- 

47 **kwargs : dict 

48 Additional key word arguments for the parent. 

49 """ 

50 super().__init__(**kwargs) 

51 self._control_model = [None] * len(self.state_names) 

52 self.control_constraint = None 

53 

54 @property 

55 def cont_fnc_lst(self): 

56 """Continuous time ODEs, not used.""" 

57 warn("Not used by this class") 

58 return [] 

59 

60 @property 

61 def control_model(self): 

62 return self._control_model 

63 

64 @control_model.setter 

65 def control_model(self, model): 

66 self._control_model = model 

67 

68 def get_state_mat( 

69 self, timestep, state, *args, u=None, ctrl_args=None, use_continuous=False 

70 ): 

71 """Returns the linearized state matrix that has been hardcoded. 

72 

73 Parameters 

74 ---------- 

75 timestep : float 

76 Current timestep. 

77 state : N x 1 numpy array 

78 current state. 

79 *args : tuple 

80 Additional arguments placeholde, not used. 

81 u : Nu x 1 numpy array, optional 

82 Control input. The default is None. 

83 ctrl_args : tuple, optional 

84 Additional agruments needed to get the input matrix. The default is 

85 None. 

86 use_continuous : bool, optional 

87 Flag indicating if the continuous time A matrix should be returned. 

88 The default is False. 

89 

90 Returns 

91 ------- 

92 N x N numpy array 

93 state transition matrix. 

94 """ 

95 x = state.ravel() 

96 

97 A = np.array( 

98 [ 

99 [0, 0, np.cos(x[3]), -x[2] * np.sin(x[3]) * self.dt], 

100 [0, 0, np.sin(x[3]), x[2] * np.cos(x[3]) * self.dt], 

101 [0, 0, 0, 0], 

102 [0, 0, 0, 0], 

103 ] 

104 ) 

105 if use_continuous: 

106 return A 

107 return np.eye(A.shape[0]) + self.dt * A 

108 

109 def get_input_mat(self, timestep, state, u, state_args=None, ctrl_args=None): 

110 """Returns the linearized input matrix. 

111 

112 Parameters 

113 ---------- 

114 timestep : float 

115 Current timestep. 

116 state : N x 1 numpy array 

117 Current state. 

118 u : 2 x 1 numpy array 

119 Current control input. 

120 state_args : tuple, optional 

121 Additional arguements needed to get the state matrix. The default 

122 is None. 

123 ctrl_args : tuple, optional 

124 Additional arguments needed to get the input matrix. The default is 

125 None. 

126 

127 Returns 

128 ------- 

129 N x 2 numpy array 

130 Input matrix. 

131 """ 

132 return np.array([[0, 0], [0, 0], [self.dt, 0], [0, self.dt]]) 

133 

134 def propagate_state(self, timestep, state, u=None, state_args=None, ctrl_args=None): 

135 """Propagates the state forward one timestep. 

136 

137 This uses the hardcoded form for the linearized state and input matrices 

138 instead of numerical integration. 

139 

140 Parameters 

141 ---------- 

142 timestep : float 

143 Current timestep. 

144 state : N x 1 numpy array 

145 current state. 

146 u : 2 x 1 numpy array, optional 

147 Control input. The default is None. 

148 state_args : tuple, optional 

149 Additional arguements needed to get the state matrix. The default 

150 is None. These are not needed. 

151 ctrl_args : tuple, optional 

152 Additional arguments needed to get the input matrix. The default is 

153 None. These are not needed. 

154 

155 Returns 

156 ------- 

157 N x 1 numpy array 

158 Next state. 

159 """ 

160 if state_args is None: 

161 state_args = () 

162 if ctrl_args is None: 

163 ctrl_args = () 

164 F = self.get_state_mat( 

165 timestep, state, *state_args, u=u, ctrl_args=ctrl_args, use_continuous=False 

166 ) 

167 _state = state.copy() 

168 _state[3] = np.mod(_state[3], 2 * np.pi) 

169 next_state = F @ _state 

170 if u is None: 

171 if self.state_constraint is not None: 

172 next_state = self.state_constraint(timestep, next_state) 

173 next_state[3] = np.mod(next_state[3], 2 * np.pi) 

174 return next_state 

175 G = self.get_input_mat( 

176 timestep, state, u, state_args=state_args, ctrl_args=ctrl_args 

177 ) 

178 if self.control_constraint is not None: 

179 next_state += G @ self.control_constraint(timestep, u.copy()) 

180 else: 

181 next_state += G @ u 

182 if self.state_constraint is not None: 

183 next_state = self.state_constraint(timestep, next_state) 

184 next_state[3] = np.mod(next_state[3], 2 * np.pi) 

185 return next_state