"""
HoD四种技术路线详细对比
"""
hod_comparison = {
    "电容感应": {
        "原理": "人体作为电介质改变电容值",
        "传感器": "柔性PCB贴合方向盘环身",
        "检测精度": "触摸/握持/离开",
        "检测延迟": "<50ms",
        "环境鲁棒": "手套↓/湿手↓/污渍影响",
        "成本": "$15-30",
        "量产代表": "Hyundai/Kia 2025+, Joyson",
        "SAE等级": "L2/L3"
    },
    "力矩传感": {
        "原理": "驾驶员微转动方向盘产生扭矩",
        "传感器": "EPS扭矩传感器(已有)",
        "检测精度": "有手/无手(二元)",
        "检测延迟": "100-300ms",
        "环境鲁棒": "不受外部环境影响",
        "成本": "$0(利用现有)",
        "量产代表": "Toyota/BMW(基础方案)",
        "SAE等级": "L2(基础)"
    },
    "视觉检测": {
        "原理": "DMS摄像头检测手/手臂位置",
        "传感器": "DMS RGB-IR摄像头",
        "检测精度": "手在方向盘/不在/位置",
        "检测延迟": "<33ms(30fps)",
        "环境鲁棒": "光照↓/遮挡↓/深色衣物↓",
        "成本": "$0(利用DMS)",
        "量产代表": "Neonode方案(OTA可加)",
        "SAE等级": "L2(辅助)"
    },
    "压力传感": {
        "原理": "方向盘内嵌压力阵列",
        "传感器": "压阻/压电传感器阵列(16-32点)",
        "检测精度": "握力分布/位置/类型",
        "检测延迟": "<10ms",
        "环境鲁棒": "不受外部影响",
        "成本": "$40-60",
        "量产代表": "Alps Alpine方案",
        "SAE等级": "L3/L4"
    }
}

print("HoD技术方案对比:")
for method, info in hod_comparison.items():
    print(f"\n{method} (成本: {info['成本']}):")
    print(f"  原理: {info['原理']}")
    print(f"  精度: {info['检测精度']}")
    print(f"  延迟: {info['检测延迟']}")
    print(f"  代表: {info['量产代表']}")

"""
电容感应方向盘布局
"""
capacitive_layout = {
    "方案A: 全环电容": {
        "位置": "方向盘环身360°覆盖",
        "传感器": "柔性PCB+导电泡沫",
        "通道": "2-4通道(分区检测)",
        "成本": "$20-30",
        "优势": "全覆盖，任意位置可检测",
        "劣势": "安装复杂"
    },
    "方案B: 把手电容": {
        "位置": "左右把手区域(9点和3点位置)",
        "传感器": "2个电容板",
        "通道": "2通道",
        "成本": "$15-20",
        "优势": "安装简单，成本较低",
        "劣势": "仅检测标准握持位置"
    },
    "方案C: 分区电容": {
        "位置": "上左/上右/下左/下右4区",
        "传感器": "4个电容板",
        "通道": "4通道",
        "成本": "$25-35",
        "优势": "位置精确，支持握持类型识别",
        "劣势": "中等复杂度"
    }
}

print("电容感应布局:")
for plan, config in capacitive_layout.items():
    print(f"\n{plan} (成本: {config['成本']}):")
    for key, val in config.items():
        print(f"  {key}: {val}")

import numpy as np

class CapacitiveHoD:
    """
    电容式离手检测信号处理器
    """
    
    def __init__(self, num_channels: int = 4):
        self.num_channels = num_channels
        # 基线校准值
        self.baseline = np.zeros(num_channels)
        # 触摸阈值 (从基线的变化量)
        self.touch_threshold = 50  # ADC units
        self.grip_threshold = 100
        # 消抖计数
        self.debounce_count = np.zeros(num_channels, dtype=int)
        self.debounce_limit = 3
        
    def calibrate(self, baseline_readings: np.ndarray):
        """基线校准（无手时）"""
        self.baseline = baseline_readings
        
    def classify(self, readings: np.ndarray) -> dict:
        """
        分类握持状态
        
        Args:
            readings: [num_channels] 原始ADC读数
            
        Returns:
            {
                'hands_on': bool,
                'grip_type': str,
                'grip_strength': float,
                'channels_active': list
            }
        """
        # 计算变化量
        delta = readings - self.baseline
        
        # 通道级判定
        channels_active = []
        for i in range(self.num_channels):
            if delta[i] > self.grip_threshold:
                channels_active.append(i)
                self.debounce_count[i] = min(self.debounce_count[i] + 1, 
                                            self.debounce_limit)
            elif delta[i] > self.touch_threshold:
                channels_active.append(i)
                self.debounce_count[i] = min(self.debounce_count[i] + 1, 
                                            self.debounce_limit)
            else:
                self.debounce_count[i] = max(self.debounce_count[i] - 1, 0)
        
        # 消抖后判定
        confirmed_active = [i for i in channels_active 
                           if self.debounce_count[i] >= self.debounce_limit]
        
        hands_on = len(confirmed_active) > 0
        
        # 握持类型识别
        grip_type = self._classify_grip(confirmed_active)
        
        # 握持强度
        if len(confirmed_active) > 0:
            grip_strength = float(np.mean(delta[confirmed_active])) / 200.0
        else:
            grip_strength = 0.0
            
        return {
            'hands_on': hands_on,
            'grip_type': grip_type,
            'grip_strength': min(grip_strength, 1.0),
            'channels_active': confirmed_active,
            'raw_delta': delta.tolist()
        }
    
    def _classify_grip(self, active_channels: list) -> str:
        """识别握持类型"""
        if len(active_channels) == 0:
            return 'none'
        elif len(active_channels) == 1:
            return 'single_hand'
        elif len(active_channels) == 2:
            if 0 in active_channels and 1 in active_channels:
                return 'left_hand'
            elif 2 in active_channels and 3 in active_channels:
                return 'right_hand'
            elif 0 in active_channels and 3 in active_channels:
                return 'cross_grip'  # 交叉握持
            else:
                return 'two_hand'
        else:
            return 'full_grip'

# 测试
if __name__ == "__main__":
    hod = CapacitiveHoD(num_channels=4)
    
    # 校准(无手时基线)
    hod.calibrate(np.array([100, 100, 100, 100], dtype=float))
    
    # 测试1: 无手
    result = hod.classify(np.array([102, 101, 98, 103]))
    print(f"无手: hands_on={result['hands_on']}, type={result['grip_type']}")
    
    # 测试2: 左手握持
    result = hod.classify(np.array([180, 170, 105, 103]))
    print(f"左手: hands_on={result['hands_on']}, type={result['grip_type']}")
    
    # 测试3: 双手握持
    result = hod.classify(np.array([175, 165, 180, 170]))
    print(f"双手: hands_on={result['hands_on']}, type={result['grip_type']}")

"""
基于DMS摄像头的视觉离手检测
参考: Neonode方案(2026.02发布)

优势: 无需额外硬件，OTA可加
"""
class VisualHoD:
    """
    视觉离手检测器
    
    利用DMS摄像头检测:
    1. 手/手臂是否在方向盘附近
    2. 手是否接触方向盘
    3. 握持类型(单手/双手)
    
    输入: DMS摄像头帧
    输出: hands_on + grip_type + confidence
    """
    
    def __init__(self):
        # 方向盘区域ROI (DMS图像中方向盘的大致位置)
        self.steering_wheel_roi = (0.2, 0.5, 0.8, 0.9)  # (x1,y1,x2,y2) 归一化
        
    def detect(self, frame) -> dict:
        """
        检测手是否在方向盘上
        
        实际实现:
        1. 手部关键点检测 (MediaPipe/YOLO)
        2. 方向盘关键点检测
        3. 手-方向盘空间关系判断
        """
        # 模拟输出
        return {
            'hands_on_wheel': True,
            'hand_positions': ['left_9oclock', 'right_3oclock'],
            'grip_type': 'two_hand',
            'confidence': 0.92,
            'latency_ms': 33  # 30fps
        }

"""
电容+视觉融合HoD
"""
class FusedHoD:
    """
    电容+视觉融合离手检测
    =======================
    
    融合策略:
    - 电容: 高精度触觉检测
    - 视觉: 手部位置追踪
    - 融合: 互补覆盖各自弱点
    """
    
    def __init__(self, cap_detector, visual_detector):
        self.cap = cap_detector
        self.visual = visual_detector
        
    def detect(self, cap_readings, dms_frame) -> dict:
        """融合检测"""
        cap_result = self.cap.classify(cap_readings)
        vis_result = self.visual.detect(dms_frame)
        
        # 决策融合
        # 电容: 手在方向盘上?
        cap_hands_on = cap_result['hands_on']
        
        # 视觉: 手在方向盘附近?
        vis_hands_near = vis_result['hands_on_wheel']
        
        # 融合逻辑
        if cap_hands_on and vis_hands_near:
            confidence = 0.98
            hands_on = True
            source = 'both_agree'
        elif cap_hands_on and not vis_hands_near:
            # 电容检测到但视觉没看到→可能是手臂搭在方向盘上
            confidence = 0.75
            hands_on = True
            source = 'cap_only'
        elif not cap_hands_on and vis_hands_near:
            # 视觉看到手在附近但电容没检测到→可能戴手套
            confidence = 0.65
            hands_on = False  # 保守判定
            source = 'vis_only'
        else:
            confidence = 0.95
            hands_on = False
            source = 'both_agree_off'
        
        return {
            'hands_on': hands_on,
            'confidence': confidence,
            'source': source,
            'cap_detail': cap_result,
            'vis_detail': vis_result
        }

"""
HoD在IMS中的角色
"""
hod_in_ims = {
    "ADAS安全基础": {
        "角色": "确认驾驶员随时可接管",
        "输入": "电容/力矩/视觉传感器",
        "输出": "hands_on: bool + grip_type: str",
        "优先级": "🔴 高"
    },
    "DMS补充": {
        "角色": "与DMS疲劳/分心数据融合",
        "示例": "手离方向盘+疲劳→紧急接管",
        "优先级": "🔴 高"
    },
    "警告分级": {
        "角色": "根据手离时间+驾驶员状态分级警告",
        "序列": "15s视觉→30s声音→60s紧急",
        "优先级": "🟡 中"
    },
    "ADAS授权": {
        "角色": "手在方向盘=允许ADAS;手离=降级ADAS",
        "示例": "L2需手在方向盘→L3可手离→L4",
        "优先级": "🔴 高"
    }
}

print("HoD在IMS中的角色:")
for role, info in hod_in_ims.items():
    print(f"\n{role}:")
    for key, val in info.items():
        print(f"  {key}: {val}")